Every week, we will release tips to help you get the most out of Mixpanel. Want to see more? Click here to see other #mpknowledgedrop articles.
Lifecycle Analysis with Cohorts
Lifecycle analysis is the process of breaking down your users into meaningful cohorts to determine which groups most significantly contribute to the growth (or shrinkage) of your user base. This kind of analysis is critical to sustaining the growth of your product.
Watch the loom video to learn more about lifecycle analysis.
Step 1. Define Activity and Usage Intervals
To answer meaningful questions with lifecycle analysis, start by determining the key event or set of events that matter to you. For example, the core function of a ride-sharing app is to book and complete a ride, so that could be a key event.
Mixpanel users often write into our support team, and they ask: what events are most important for me to track? The answer is inevitably it depends. What is the core function of your app? What do you want users to do? How do make money? Ask yourself these questions to figure out what your key event should be. It should feel straightforward — there’s no complicated grad school formula here.
Once you’ve identified a key event that you are interested in analyzing, determine what is an appropriate usage interval. That’s the natural frequency at which your customers use your product. It could be daily, weekly, monthly, etc. If you’re not sure what this is, use Mixpanel’s retention report to generate a “Power User Curve” to understand the natural usage frequency of your key event or do a cumulative distribution frequency function.
Users that perform your key event within your defined usage interval can be considered as active.
Step 2. Define Cohorts
In lifecycle analysis, the most commonly used cohorts are new users, resurrected users, retained users, and dormant users. The following table shows what each cohort represents. In your product’s use case, “Did action” would be replaced by your key event.
The next screenshot is an example of the Retained User cohort, where the key event is called “Message Sent” and the usage interval is 7 days. Make sure to select rolling range, so the dates of the cohort criteria adjust based on the current day. Since this is the retained user cohort, according to the table above, I specified that I want to include users who performed “Message Sent” in the current period (last 7 days) and also the previous period (14 to 7 days ago).
Once you’ve created your 4 cohorts, you can visualize them over time in Insights to track how each group of users is changing over time. Expect to see something like this:
Now that you have this graph of your most significant cohorts, you can ask questions to understand which user groups are impacting your growth and make decisions for how to influence growth.
Check out more examples of questions that you can answer with lifecycle analysis on our blog.
Measure Your Product’s Quick Ratio
Quick Ratio is the ratio of users added to your user base over users lost. Users added is made up of new and resurrected users. Therefore, you can use the cohorts that you created for lifecycle analysis to plot this metric. Quick ratio determines whether your net user base is growing or shrinking. If the ratio is above 1, then your user base is growing. If it is below 1, then your user base is shrinking.
Track Movement of Users Between Cohorts
Define significant cohorts
A significant cohort is a group of users that have performed an event or share a property value that matters to the success of your product.
In this sample use case, an “occasional watcher” is a user that performs Watched Video 1 or 2 times in a 7 days period. A “core watcher” is a user that performs Watched Video between 3 and 10 times in a 7 days period. Finally, a “power watcher” is a user who performs Watched Video 11 or more times in a 7 day period.
The following screenshots show how to define cohorts that capture the number of users that move from the “Occasional watchers” and “Core watchers” group to the “Power watcher” category.
Occasional to Power: the number of users that were in the Occasional watchers group 14 to 8 days in the past but moved to Power watchers in the past 7 days.
Core to Power: the number of users that were in the Core watchers group 14 to 8 days in the past but moved to Power watchers in the past 7 days.
Visualize Cohorts in Insights and Explore
Now you can go to Insights and plot these cohort trends over time. You can plot these two cohort trends as a dashboard card or set an alert if you want to track this as a key metric.
You can also go to Explore and view the exact set of users that moved between two cohorts, and click into specific user profiles.
You can even use these cohorts to send messages to users when they move from one group to another — but that’s for a different post!