Organizing a data team
AA: Speaking of teams, some organizations ask themselves, “Should we be centralizing the data science function? Should it be distributed and based in different business units?” Do you have a preference that you’ve used in the past? Of course, it depends on the organization and the analytics maturity, but do you tend to favor one over the other with how you structure the teams that you manage?
AG: When you’re first starting out, if you’re a small company, you can’t have a lot of different teams. Generally, you start with a centralized approach. Because the team is centralized, all the data is together; you have one source of truth. And generally, those people are good with the technical areas. They can solve technical problems quickly, and they can talk to each other. Knowledge sharing within the data team is very good. It definitely has a lot of benefits.
The challenge is that domain knowledge is also very important...