User behaviour analytics 101

User behaviour analytics 101

New products rarely get the magic founders to imagine. This article demonstrates how founders can leverage data to build a more user-centered product.

This article is mainly focused on phases after product release. I am addressing it to founders and assuming that by now you have orchestrated your series of discovery meetings with customers. All signs have led towards creating your current product version.

Unless you are a marketing genius new products rarely get the magic founders imagine they would. Everyone is faced with a cold start problem and my goal is to help you see through some decisions you will need to make in order to reach the next phase. For example, if you thought you will get 1000 users to register in the first week it would probably be 100 and from that 100 you will probably lose 80% due to the bad user experience.

I will be using the Mixpanel tool to demonstrate how you can leverage data to build a more user-centered product, tweak your marketing strategy, and track growth. You can read more about Mixpanel here. The following can also be done with some other service. I am using Mixpanel since our team had success using it for upselling services, driving renewals, reducing churn, and increasing engagement.

How to track user behavior?

With correct implementation, you will be able to determine if something is broken and with the wrong one you will be stuck with another useless subscription to a service you don’t need. In other words, you don’t need to know how many times users have clicked on button X, but it helps to know where most users decide to close the app, at what step they cancel the onboarding process, what cohort they belong to, and what they can’t live without.

To do it properly we need to take a step back. We need to ask ourselves which KPI Owner wants to measure user behavior. For example, look at the table below:

Screenshot 2022-07-19 at 15.40.23.png

For example, let’s say that we want to measure onboarding funnel conversion, since this is usually the first pain point founders come across (Ask for email?, require a phone?, use SSO?, do a walkthrough?). We will be building an influencing app where influencers are able to share their content and get paid by followers on a monthly subscription basis. We want to learn how many influencers were able to onboard to our app and if there are any obstacles that we can fix.

The formula for a good tracking plan could be creaked down to:

  1. Goal: Something you are trying to accomplish
  2. Questions: How are required input fields affecting onboarding percentage?
  3. User flow: Chain of events/ user behavior you need to look into in order to get an answer. All events need to have certain properties attached like age, type of user, etc.

In our example:

  1. Goal: We are trying to increase the onboarding percentage of influencers. Both growth and product owners are usually interested in these numbers.
  2. Question: In our example, we want to learn if any extra data (birthdate, phone, username, location) presents a bottleneck. Are some of these required fields increasing bounce rate?
  3. User flow: Influencer needs to:
    • Create an account (phone, email, Google, or Twitter)
    • Confirm account (optional step)
    • Enter extra data (name, phone, username & birthdate)

Once we have our tracking plan we can build out a funnel in Mixpanel. An example would look something like this

Group 3.jpg

Understanding the reports

With the following report, we are now able to learn that we have a bottleneck in the step where we are asking for extra data. With just one additional breakdown based on entered data, we should be able to optimize onboarding flow and eventually get a higher conversion rate.

Group 3 Copy.jpg

Similar reports and later even the whole dashboard could be constructed for different personas that have questions about product performance. Like we like to say products are living things there is always room for improvement and if that improvement can be calculated out I will always prefer it this way.

Hope that you enjoy this short article and it will give you a better idea of how products can be improved with a little help from data. Have some questions? Drop me a line at and I’ll do my best to answer.

Stay in the loop


Welcome to our email list

We promise we won't spam you!

Marko Balazic

Product Design