Data Tips #20 - Organisation: Roles in a data team

Starting off a series of organisation articles to discuss the different roles you may need in your organisation to become datadriven. As always, which roles you need depends on your company and what you want to achieve.

Let us start with the roles for the data team. In general I think the data team should be able to handle end-to-end use cases: from sourcing data and transforming it, to analysing and presenting it. There are many surrounding roles needed to be successful, and I will cover those in upcoming articles.

The responsibility of the data team is to execute on use cases and implement the needed functionality. So what kind of people do you need on the team? And so I do not repeat myself for each role: all of these roles work together and communicate together. There is no sequence and there are no handoffs. Everyone is in the car for the ride.

Business translator. We need to start at the beginning: what is the business problem you are trying to solve? The role of the business translator is to facilitate the understanding of what the use case actually needs, working closely with business experts, engineers and architects to make sure the goal is clear.

Data engineer. Makes data available, reliable and reusable - pipelines, models and the operational quality of both.

Analyst / data scientist. Turns available data into an answer, a model or a decision support that someone actually uses.

Analytics engineer / modeller. Owns the semantics: the definitions everybody argues about once and then stops arguing about.

Staff the team so a use case can go from question to production without leaving it. That single property explains most of the difference in delivery speed I see between organisations.