Let’s face it. Most product back-end databases weren’t designed for ease of analytics – they’re built to efficiently power software. Easily answering curious, thoughtful and complex analytic questions from them isn’t the priority of these databases – but analysts still have a job to do. That’s why Product Database Literacy is the killer skill for all aspiring Analytics Engineers.
Here are our 5 steps to becoming an expert on any product database:
To start, identify the obviously important tables. Take a look at them. Explore their data.
Ask questions about how the app uses this table and then run queries to test those assumptions.
Next, spread out a bit. Notice the foreign keys in those core tables that seem relevant to how the app behaves. Find the primary tables for those foreign keys and investigate those tables too.
If you can, get access to the front-end product and run tests comparing front-end user activity to back-end data.
Keep repeating until you have a logical understanding of how the app behaves, based on its back-end data. It might be useful to draw out the important tables, map their related keys and list any columns particularly important for answering analytic questions.
Then start building a few simple reports and summary answers to commonly asked questions. Take a step back and ask, do these results make business sense.
Ask someone who isn’t an Analytics Engineer if the results make sense. You’ll probably learn a bunch of things about the subtleties of how the data works, especially with edge cases.
Update your data model. Write down your key learnings and queries that explain how the product database works. And don't throw away important queries! You spent time sculpting those during your investigation, so make sure to save them somewhere.
Finally, walk someone else through it too. They’ll ask questions and harden your understanding of the app database. Soon, there’s no question you can’t answer!
You’ve just performed database archeology and you’re now an expert. Take a second to high-five yourself.
In just a few hours, you’ve gone from product database novice to expert! You’re now the company expert on this data.
Drop us your thoughts in the comments or connect with us to chat more!