Prerequisite: Guide to Mixpanel Basics
This guide assumes that you already have a foundational understanding of the Mixpanel data model. If you haven't already done so, please take time to review the Guide to Mixpanel Basics before proceeding with this guide.
Mixpanel offers an array of analysis tools that provide actionable insight into both a product and its users. In this guide, we'll explore a few of Mixpanel's most popular analysis features and discuss important best practices to consider when analyzing your data.
Understand Your Data
Before you try to run any sort of analysis, it's important to familiarize yourself with the data that is being collected within your Mixpanel project. The easiest way to do this is by navigating over to Lexicon which contains a complete list of every event and property that is sent to your project. Here, you can also find additional bits of information on each event and property to help you better understand the data in your project.
If you come across an event or property with a missing description in Lexicon, it simply means that no one at your company has set a description for that event or property yet. In such cases, try to track down the individual(s) responsible for setting up your Mixpanel tracking and ask them to add the appropriate descriptions. Only user accounts with either an Owner or Admin project role will have the necessary permission level for adding descriptions within Lexicon.
This same information about events and properties can also be accessed directly within any of the Mixpanel reports. Hovering over an event or property within any dropdown in Mixpanel will cause a panel to appear with the Lexicon details about the corresponding event or property.
Adding descriptions and displaying volume and query counts are just a few of the features offered by Lexicon. Check out this article for a complete overview of Lexicon's functionality.
In this section, we'll identify some common product metrics and see how easy it is to analyze them with Mixpanel. Through this process, you'll be introduced to a few of the most popular analysis reports and features within Mixpanel.
We'll be using a hypothetical music streaming app called Music Finder as the context for all upcoming examples in this guide. While the data shown in these reports will be specific to Music Finder, all concepts illustrated in these examples are universally applicable, regardless of product or industry.
Number Of Active Users
Measuring the number of active users over time is a common way of analyzing the growth of a product's user base. While this is a fairly universal metric, the definition of an "active user" will vary from company to company – let's look at a couple examples.
The most basic way to define an active user is by identifying a single event that the user must perform in order to be considered active. For example, let's consider any user that has played a song in Music Finder as an active user. In that case, we can simply use the Insights report to show us the unique number of "Song Play" events as a line chart over time.
By adjusting the time interval, we can switch between analyzing the number of daily, weekly, or monthly active users.
Alternatively, we can use Insight's built-in DAU, WAU, or MAU function as a shortcut for building our active user report.
Calculating the number of active users is just one of many types of analysis that can be performed with the Insights report. Check out this article to learn more about what's possible with the Insights report.
Let's say we change our minds about how we want to define an active user – now we only want to consider a user as active if they have played at least 3 songs. To calculate this more advanced active user metric, we would first need to use the Cohorts feature within Mixpanel to define a group of currently active users.
Next, we can view this cohort within the Insights report to see how the size of our active user cohort has changed over time.
Cohorts is an extremely powerful tool that allows you to define groups of users based off of advanced filtering criteria. Once a cohort is defined, it can be used across the various Mixpanel reports to analyze the behavior of a particular subset of users. Check out this article for a complete overview of Cohorts.
Conversion rates are a measure of how many users successfully move their way through a defined user flow. The Funnels report makes it extremely easy to analyze conversion rates – let's look at a quick example.
Let's say we want to measure how many users go on to sign up for an account after viewing our signup page. In that case, we can build a simple two-step funnel to measure our signup conversion rate.
Furthermore, we can view the conversion rate as a trend line to see how it has changed over time. This feature is particularly useful when trying to assess the impact of product changes on important conversion rates.
The Funnels report supports a variety of different configuration options which will impact the way the analysis is calculated. Check out this article to learn more.
One of the best indicators of a product's health is its user retention rate, which measures how effective the product is at keeping users coming back over time. The Retention report offers a simple interface for measuring various types of user retention rates – let's check out a couple examples.
Select either one event or two different events. Retention with one event measures how many users repeatedly perform the same action over a period of time. For example, let's say we want to know how many users that were playing songs in Music Finder last month are still playing songs in our product this month. We can easily analyze this with a Retention report.
Retention with two events includes one event to create the user bucket you are measuring, and another event to measure the retention on. For example, let's say we want to how many users performed a specific event (such as sign-up), and how often they came back to do some other action on your site.
Please refer to this article for details about the nuances of how the numbers within a Retention report are calculated.
Properties in Analysis
The value of collecting details about users and their actions in the form of properties is that these properties can be used to perform more granular analyses within Mixpanel. More specifically, there are two primary ways that properties can be utilized within Mixpanel, and we'll explore both in this section.
Property filters allow us to narrow the scope of our analysis to a subset of data that meets particular criteria. For example, let's say we want to know what our signup conversion rate is for users within the United States on the Firefox browser. We can get our answer by running a Funnels report with a couple property filters.
Property breakdowns allow us to compare different subsets of data side-by-side. When we select a property breakdown, Mixpanel will first segment our data into groups based on the value of the selected property and then perform the specified analysis across each group separately. For example, if we wanted to know what musical genre has been the most played this month, we can accomplish this with a property breakdown in an Insights report.
Furthermore, most Mixpanel reports support multiple property breakdowns. For example, if we want to know what musical genre is most popular within each country, we could first breakdown by the "Country" property and then breakdown by the "Genre" property.
Event Properties vs. User Profile Properties
Mixpanel allows you to use both event properties and user profile properties within filters and breakdowns. It's important to know the difference between these two so that you can decide which one is appropriate for your intended analysis.
Event properties provide details about an action at the moment it took place whereas user profile properties describe the current state of a user. Depending on your implementation, you might have instances where an event property shares the same name as a user profile property. For example, in our Music Finder project, all events have a "City" property (i.e. event property) and all user profiles also have a "City" property (i.e. user profile property).
The values for these two properties can differ, even though they refer to the same user and share the same property name. For example, if a user played a song in San Francisco yesterday and then logged in to Music Finder from Los Angeles today, the "Song Play" event from yesterday would have a "City" property equal to "San Francisco" but the user's profile would have a "City" property equal to "Los Angeles."
In this guide, we only scratched the surface of the various types of analysis that can be performed within Mixpanel. For a more exhaustive list of Mixpanel's analysis capabilities, please explore this collection of articles.