Action Cohort Analysis: A key tool for understanding user engagement
So you’ve been tasked with “monitoring user engagement” in your mobile app. But wait a second, what actually defines good user engagement? “Engagement” can be interpreted in many ways and applied differently across mobile app categories.
As Intercom elaborates, “For a to-do app an engaged user should be logging in every day to add and complete items whereas for an invoicing app, an engaged user might only log in once per month. There is no consistent quantifiable definition of engagement across different products.”
Despite the fact that there is no quantifiable definition of engagement across varying verticals, there are still some basic metrics that mobile professionals gravitate towards. With any standard analytics tool, you’ll probably start by tracking metrics such as daily app launches, churn rate, and daily sessions per DAU.
These are important metrics, but there is so much more that you can do for measuring user engagement.
There’s one tool in particular that can provide you with a deeper understanding of user engagement: action cohort analysis.
Action Cohort Analysis: What is it?
“Cohort” is a term that’s used to describe a group of people who’ve banded together due to the same attribute. For example people born between the years 1980 to 1990 who have attended a private university are a cohort. In the case of mobile app cohorts, a cohort is a group of users who have performed a common action or actions during a specific time frame.
In app analytics, action cohort analysis allows you to create groups of users based on what they do in your app within a specific time frame (whether that be a day, week, month etc.). Action cohort analysis can help you understand how users are actually interacting with your app and examine trends in behaviors over periods of time. By enabling you to view your users as specific, defined groups, rather than a single aggregate unit, you can better identify engagement and usage patterns throughout the user lifecycle. These patterns can help direct you towards the data that really matters and allows you to make more informed decisions on your app’s optimization.
To help you better understand how cohort analysis works, let’s first go over the actual setup of this tool. When creating an action cohort you need to set up three parameters.
Now that you have defined your time frame, first action, and section action, you are good to go. If you have defined the time frame as a current month or week, you will need to wait for your behavioral analytics tool to populate with enough data. Bear in mind that in cohort analysis, a user will only be counted once per cohort, but can be included in more than one cohort. For example, let’s say you have a weekly cohort that is based on the first action of a completed ‘login’ event. That means that a user who actively logged in at least once each week will be in every cohort, and not only the cohort for their initial login.
Take a look at the visualization below of an action cohort based on a weekly time bucket. Notice how the same individual is not repeated within the same cohort but can appear across multiple cohorts depending on your initial selected action. If you have defined the initial action as ‘started 1st session’, then you would not see that individual repeat across cohorts as that action can only occur once.
Pro Tip: Cohort analysis + user session recordings
Action cohort analysis is a great way for you to ask a specific question about user behavior and/or usage trends, and then examine relevant data. However, in order to confidently act on this data, you need to try to understand why you are even seeing those numbers. A key question to ask here would be why some of your users are not completing a second action (and trust us, there will be at least “some”). A lot of hypothesizing and “educated guessing” is likely involved. Today’s users are complex and it can be challenging to apply one reason for an incomplete action.
There is an effective work around that can save you time, resources, and stress. It comes in the form of user session recordings.
Mobile app analytics tools that provide user session recordings, like Appsee, enable mobile professionals to watch session recordings of users within specific action cohorts.
For example, let’s say you set up an action cohort with the initial defining action as ‘Started First Session’ and the second action as ‘Triggered Event: PurchaseComplete’. With qualitative analytics tools, you can watch session recordings of users who did not complete the second action within a certain time frame. Maybe there’s a crash happening for some users who attempt to pay via Paypal? Perhaps there is a confusing text element within the billing vs. shipping address section? The answers to these questions can be visualized via the pairing of session recordings with cohort analysis.
How does this combo apply to other app categories? Let’s take a look at some powerful use cases.
Business/Productivity
So you have a business app that enables users to manage their tasks. A powerful action cohort for your app could be to track how many users started a session and then came back and tapped the “+” button to add a new task. This cohort will help you examine whether users are adding a new task to their task list as frequent as you anticipate. With user session recordings it will also enable you to see what friction points might have stopped them from adding the task.
Gaming
If you’ve got a gaming app and you want to optimize IAPs, you need to be able to examine your users’ in-app purchase patterns. A good way to do this is via an action cohort with the first action as an in-app purchase (i.e. buying coins) and the second action as an in-app purchase. This can help you understand on a daily, weekly, and monthly basis, how quickly and often your users are willing to make recurring purchases.
Entertainment
Just as tv channels want to understand how many users are tuning into episodes of a show, an entertainment app should want to analyze how many users are streaming videos. In this case, you can create an action cohort with the first action as “Started a Session” and the second action as “Play Video”. This way you can dive into the how many users are streaming videos (or specific videos) after starting a session and how frequent.
Health & Fitness
Most Health & Fitness apps have an onboarding process that requires users to complete specific weight, diet, height, and/or exercise preferences. While this information is typically necessary for the user to complete, only “committed” users will actually follow through with this requirement and then use the app. To identify the commitment level of your users, you can set up an action cohort with the first action as “Completed Onboarding” and the second action as “Started a Session”. You can then watch session recordings of users who did not complete the onboarding process and zoom in on the exact reasons as to why.
Bear in mind the key to effective action cohort analysis is composed of two parts.
- Selecting the right parameters for actions and behaviors relevant to your app
- Utilizing a powerful behavioral analytics tool that provides insights beyond quantitative metrics
Can’t get enough of cohort analysis? Check out this live-action visualization of action cohort analysis via this tutorial.
Better yet if you want to see action cohorts on your own mobile app, you can easily access cohort analysis (plus session recordings) via a 14-day free trial with Appsee.
