While many teams scratch only the surface, Mixpanel offers deep analytical capabilities that can revolutionize your data-driven decision-making. We'll explore features like action correlation to purchases and root cause analysis for KPI fluctuations.

‍

Uncover Actions that Drive Purchases

Many teams underutilize Mixpanel's Apps section, dismissing them as mere side features. However, these tools are potent for data analysis.

The Signal feature allows you to identify actions that correlate with specific user behaviors. While retention is the default focus, you can also analyze actions that lead to purchases.

First, create a cohort of users who have made a purchase in the past 30 days. After you’ve saved the desired cohort, select it as the Goal. Make sure the "For" part is targeted to your pre-goal group of users as well. In my case, I want to see what actions correlate with New Users making a purchase, in the last 30 days.

‍

Then, voila! You have yourself a purchase correlation table. The stronger the correlation score, the more the action is correlated to making a purchase.

As a side note, there are two things to be careful of when looking at correlation scores.

First, any correlated events over 0.8> are usually just noisy events that are pre-requisite actions before a purchase. These are actions like “add to cart”, or “confirm purchase”. They show up as highly correlated, simply because they are actions the users are required to perform in order to make the purchase in the first place.

Secondly, we need to be cautious of high correlation scores attached to really rare actions. For example, I've seen the action "change advanced settings" in an app have a high correlation for new users becoming paid users, but only 0.9% of new users were even taking that action. At first glance, we might interpret this result as indicating value in changing advanced settings for new users. Consequently, we might consider directing more new users to the advanced settings page or incorporating it into the onboarding flow. However, the reality is that those 0.9% of new users who were changing their advanced settings and then becoming customers were not actually new users; they were veteran users with new phones, leading our analytics to misclassify them as new users. Naturally, these veteran users would be familiar with advanced settings that most new users wouldn't care about, and they would be more likely to convert to paid users than the average new user. That's why, when we see a high correlation score attached to an extremely rare event, we need to investigate further.

Root Cause Analysis for KPI Fluctuations

Have you ever scrutinized an important KPI, frustrated while trying to determine why the numbers went up or down? Mixpanel actually has a hidden root cause analysis feature that you've likely overlooked numerous times. For any conversion funnel or retention chart, there is a "Find Interesting Segments" button found at the bottom of the chart.

If you click it, a few minutes later, Mixpanel will send you an in-depth report to your email, on the user segment that is likely causing the change in the chart’s numbers. It’s almost like magic.

You can even click into each of the segments, and it’ll show you the exact segment they’ve targeted.

Mixpanel is more than a basic analytics tool; it's a treasure trove of advanced features waiting to be explored. I hope these tips helped some key features you may have overlooked, empowering you to make more informed decisions. If you want to dive deeper into Mixpanel and unlock its full potential, claim your free consultation with Shoka Sonjuku, we're happy to help.