Don't rely on your engineers to decide what data to send to Mixpanel, do this instead!

  • 24 July 2020
  • 2 replies


Hey there :wave:

I'm a happy Mixpanel customer (I lead growth at Integromat) and I want to share how we went about our Mixpanel implementation. Unlike most companies, we did not start sending data to Mixpanel right after purchasing the license. In fact, it took us upwards of 2 months to send the first set of events to our production project on Mixpanel.

So what took so long, you ask? :thinking:

Well, we spent a ridiculous amount of time creating our tracking plan which, as most of you know, is the source of truth of all your customer data.

One of the common mistakes people make is to start sending data right off the bat only to realize that they need to modify the taxonomy or simply don't need some of the events and properties being sent.
This not only results in wasted engineering resources, but also makes your Mixpanel projects bloated with unnecessary data, making everything harder from that point on.

In our case, we spent a fair amount of time coming up with the burning questions we wished to answer, and based on that, defined the first set of events that we wanted to see on Mixpanel. Alongside, we narrowed down on the requisite entity data (user and group properties) that were crucial for us to understand user personas and tie their activity to their respective accounts.

It might seem a bit much spending upwards of 2 months on the implementation but in hindsight, it has been the best decision. Every piece of data that is being sent to Mixpanel is named and described properly, making it easy for folks on other teams (marketing, partnerships, customer support) to make sense of what they see.

To sum it up, don't underestimate the value of a proper tracking plan — it can quite literally make all the difference when it comes to gathering clean and consistent data.

At the end of it all, I realized that there is a lack of in-depth learning material on this very important topic of creating tracking plans. Moreover, there is little to nothing that covers the ins and outs of Customer Data, enabling non-engineers to lead the implementation of event-driven analytics and engagement tools.

Therefore, I ended up creating a course in a format that is easy to digest and enables rapid learning.

Check it out here:

I’m happy to answer any questions you might have about creating tracking plans or deciding what data to track when you’re getting started! 

2 replies

Userlevel 5
Badge +3

@ebrawer — I noticed your post around naming best practices, and thought you might find

@arpitc’s resources helpful. He has a free module specifically on data types and I bet could take a look at your question too. 


Thank you @cherise for the shoutout! :raised_hands: