Truly understanding our customers and their problems is the most important job of a product manager. When a new feature is released, the most pressing question is: “Are users using it? How do I use it?” To see if we made the right assumptions and created what the user wanted.

Metrics like daily active users don’t answer these questions. To truly understand the impact of my decisions on the users of my product requires a deeper level of analysis — user behavior analysis.

Customer-centric teams need to be acutely aware of user behavior:

  • Fully understand the needs of users
  • Understand what features customers use and what they don’t use
  • Understand how customers can get the most value out of their products

What is user behavior analysis? Why is it so important?

User behavior analytics are data on how customers behave in mobile applications or websites. It goes beyond basic metrics like monthly active users or aggregate page views. Behavioral data reveals how interactions with products affect retention, conversion, revenue, and the results you care about.

Understanding user behavior is essential to increasing engagement, retention, lifecycle value, conversion rates, and ultimately revenue.

Behavior analysis is based on events. Events represent any action a user can perform within the product (such as opening an application, creating an account, watching a video) or any activity associated with the user (such as making a purchase).

Sending event data to the analytics platform is the most important step in understanding how users interact with the product. If you rush your analysis too quickly, you may never get the full value of your data.

Follow these 10 steps to start analyzing user behavior the right way.

Step 1: Define business and analysis goals.

What are you and your team doing? What is the overall goal? Think about the business goals to be achieved. Once the overall goal is established, how will it be achieved? Establish key performance indicators (KPIs) that focus on improvement to achieve their goals.

Assuming the business goal is to increase revenue, the KPIs might be:

  • Increase retention of paying customers
  • Increase payment channel conversion rate

Before you start thinking about data taxonomies, you must define them to ensure that the right events can be sent to the right projects to track KPIs.

Step 2: Find the critical path that matches your goals.

The critical path is a set of actions that users take based on the purpose of the product. Examples of e-commerce products might be:

Search → Browse products → Add to cart → Checkout → Order confirmation

For a game product, the critical path might start with the user opening the application, prompting for registration, and then following through the game’s tutorial.

Example: Suppose you have a game application. A critical path for the application can be broken down into four different events: the user opens the application >> User registers >> User validates the account >> User completes the tutorial

To start, be sure to track only events that are critical to answering business and analysis goals. You can always add more events later if needed.

Step 3: Organize activity categories.

Behind every good user behavior analysis is a good event taxonomy — a way of organizing a collection of events and attributes that define the actions that people can take in the product. It is critical to get the event taxonomy right as the basis for all future analysis that will be performed using the analysis platform.

If you are not using a product analysis platform to organize event categories, you may use the following spreadsheet to track all the names of events, event attributes, and user attributes.

Step 4: Learn how to identify users.

Most analytics platforms require some sort of identifier (for example, user name or email) to be configured in their mobile SDK or HTTP API to keep track of unique users. This allows you to match data from multiple devices and sessions with a single user. Therefore, it is important to ensure that the user ID is included.

Another important thing to note. Most analytics platforms count uniquely identified users when they “see” a new device or a new user ID (if the user is logged in). The problem arises when the device “anonymously” records events that are actually executed by users already in the system.

Step 5: Determine whether cross-platform behavior analysis is required.

If your product exists on multiple platforms, such as mobile and PC, or in-store POS and native applications, should you bundle all your data together or store it separately?

The answer to that question depends on the product. If you expect different user behavior across platforms, you need to understand how each platform performs independently of the others, so cross-platform detection may not be a priority.

If you want to understand the entire user behavior across the entire user path, make sure the analytics solution can detect across platforms.

Instacart, a grocery shopping app, is a great example of a product that leverages cross-platform tracking because they want to know how people are using their product on the move and on the web.

Step 6: Establish minimum viable procedures.

Once you’ve spent some time thinking about how to set up analysis and organization events (aka steps 1 through 5), it’s time to start accessing some basic application metrics. Now you should integrate the analytics solution’s mobile SDK and/or HTTP API and assign user ids.

Step 7: Track events.

Start tracing the events and critical paths in Step 2. It is not necessary to track every action that may occur in your application, but make sure that tracing events is a key step in pulling new, transforming, and preserving.

Step 8: Set user properties and event properties.

Assigning user attributes and event attributes gives you more insight into the behavior of customers when they interact with your application.

  • User attributes Describe personal attributes (such as age, gender, location) that use the application.
  • Event properties Properties that describe an event (for example, when someone executed the event)

Step 9: Verify that user behavior events are properly tracked.

How do you know if you’ve tested everything correctly? Browse the application using the test device. If you can view the analysis in real time, you should be able to see device trigger events at each step. Once the test data is validated, it’s time to start sending live data.

Step 10: Start digging into user behavior.

Recording user behavior events is a huge achievement. This is a meaningful investment that gives the team access to the data needed to assess the impact of product decisions.

When you’ve done enough testing, it’s time to start using all of this data.

  • Building behavioral groups
  • View the critical path and channel reporting to improve conversion rates
  • Calculate the user retention rate over a period of time
  • experiment
  • Measure the impact of new feature releases
  • Generate more insights