5 Minute Guide to Product Analytics

I get a lot of questions like "how to use analytics" and "what to track", and wrote this quick guide for a friend. This is specifically for "product analytics", which I thought might be useful to folks here. The thoughts here come from questions and lessons from my experience at Mixpanel.

Product Analytics boil down to two categories...

1. Metrics. With metrics, you want them to go up or down.

2. Conversion. Conversions are important because you get an exponential return. A 1% bump in a Sign Up flow conversion, is better than a 1% bump in how many folks see the first page.

To integrate analytics, I'd recommend two things:

1. Write down the business questions you want to answer with your data
2. Track the fewest amount of events possible, with many properties

TLDR; rules of thumb...

• track no more than 10 events if you're just starting out
• use generic events like: Page View and Purchase
• track as many properties as you want (attach information to these events)

The simplicity of events will make it easy to find answers to basic questions. The addition of extra information on these events (properties) will allow you to slice and dice the data later on. It will give you access to more complex answers and help you improve your metrics or your conversions.

Tracking Plan for a standard E-commerce use case (event name, then properties)

• Page View , name = landing
• Searched, query = winter coat
• Page View , product = mens winter coat
• Add to Cart, product = mens winter coat , size = medium, cost = ... etc.
• Purchase, total amount = ..., items in cart = ..., shipping cost = ... etc.

Using these four events, we can answer a lot of questions. What's the average price for the order? What's the conversion of our purchase funnel? How many people saw our home page?

Then you take it a step further, and ask, "How do we make it faster for someone to checkout?". You can dig deeper into your data, experiment with solutions, then see if your solutions had an impact on your metrics or conversion.

Hope some of this was helpful for the group, here's a link to the full article if you need even more detail. Let me know if I've missed anything or any questions come up!