So true. In my last company, the company that hiring a VP of Data and throwing money at looker, heap and snowflake would make them data driven and it didn't.

While I agree with a lot of things mentioned here, like data team being more proactive and getting to know business, you always run into a scaling problem. There are only so many people you can hire.

Being data-driven means that you need the data ethos to be dispersed across the company. Your PM's, your engineers and TL's, all of them need to "understand" data and be able to get their hands dirty. This was the culture I built when I ran a data-driven team as well as when I ran data analytics teams.

To that point, I liked your semantic layer concept. You need a team of specialists who can get the heavy lifting out of the way, then you empower feature teams to do the data interpretation and whenever possible customizations.