In Q1 we want to Ship an MVP of Automated Insights to validate this as a long-term direction.

@EDsCODE and I have been talking a bit about this but I wanted to get it in a public thread so we can continue the discussion and get feedback from others.

What is automated insights?
Previously called "Universal Taxonomy", Automated Insights is about answering the important questions our users haven't even thought of asking yet. We'll make this possible by understanding their events, e.g. their "Add to my basket" is the same as our "Add to Cart" and we know that it's common to show use a funnel with add to cart to understand conversion rates on an e-commerce app - so we suggest this to a user. Demo

What's the objective?
The goal of this project is that we validate (or invalidate) this as a long-term direction by de-risking the following:

• Is it possible to create a taxonomy that covers >80% of events for multiple businesses (for a single vertical e.g. e-commerce)
• Can we automatically map real customers events to this taxonomy
• Can we suggest useful insights to real customers before they think of them

Whats the scope of the MVP?

• A single vertical (e.g. e-commerce or healthcare)
• ~5 of customers
• End-to-end experience (not just a spreadsheet or query)

What should our next steps be?

• Identify the best vertical to work with, based on customer coverage and ease of working with events
• Identify customers to partner with on the solution
• Create first version of taxonomy for vertical
• ...

Aren't we avoiding building low quality partial solutions?

• We need to take some risks for the long term to stay ahead of our competitors and stop our product from becoming stale
• If this is successful - it will enable us to create am even higher quality (easier) experience for a subset of customers
• If this fails we will not keep the code lying around, ensuring we learn from it but don't reduce overall quality

Would really appreciate your feedback and further thoughts on this.