I spend a lot of my time talking to the heads of data science teams about the projects they are working on and the data they need to complete their work successfully. Recently, I spoke with a manager who had just joined his firm and questioned how his team is currently organized. He felt like the structure and culture were holding the team back.

That conversation sparked me to speak with many more managers at various types and sizes of firms. I asked them how they organize their data science teams and how they divide up the work that needs to be done. Here’s what they said…

You Hired Smart People – Let Them Be Smart.

Almost every manager I spoke with admitted they struggle at times balancing the goals of the firm and the need to let their data scientists do science. Top-notch data scientists need to be allowed to ask questions, try out ideas, learn from their failures, and not be micromanaged from above.

Too often, organizations assign a person to a particular area or task, and discourage him/her from investigating anything else. Smart data scientists need to be allowed to look at whatever areas or problems they find interesting. This fosters a more creative and energetic data science team.

Abandon the Silo and Collaborate.

Several managers told me they had worked in both siloed and collaborative environments. Every one of them preferred a collaborative environment. Very hard problems are rarely solved by one person. Bringing in others with diverse opinions and expertise in particular areas can lead to a better solution in less time.

The managers prefer to divide up complex problems into smaller, more manageable pieces and assign each piece to the person best suited to solve it. If a team member is hiding in the corner, working on only-he-knows-what, and not talking to his colleagues, you have a problem.

Work in a Lab, Not an Office.

Another characteristic of a successful data science team that was shared with me is fostering a science-based environment. Decisions should be made using scientific methods. The culture is less business-oriented and more academic. Such a culture fosters debate, holding peers to a high standard, and reaching common conclusions about research questions.

Great People Deserve Great Tools.

Finally, all the managers I spoke with said it takes more than just hiring the smartest people, nurturing a collaborative approach to solving problems, and creating an academic environment. They also need the right tools to get their work done. So these managers make sure their teams have the computers, software, and data they need.

The Takeaway.

There appears to be a consensus amongst successful managers on how to organize a data science team:

  • Allow your people to be curious and free to explore topics of interest.
  • Collaborative environments yield better results than sitting in silos.
  • Foster a more academic setting that uses scientific methods to solve problems.
  • Give your people the tools they need to do the hard work.

What about you? Do you have any “dos or don’ts” for how to organize a data science team? Please comment – I’d love to hear your thoughts.

Also, please connect with DIH on LinkedIn and Twitter.

Thanks,
Tom Myers