Structuring Your Data Team: 9 Best Practices
Hire or grow from within? Structuring a data team isn't easy because there's no one way to do it right. Here's a look at some pitfalls and best practices.
![](https://eu-images.contentstack.com/v3/assets/blt69509c9116440be8/bltf602e72ee2cfb3cf/64cb1662784c19295ca1afda/football-622873_1280.jpg?width=700&auto=webp&quality=80&disable=upscale)
Structuring a solid data team would be a lot easier if there were a common blueprint that worked equally well for all organizations. However, since each organization is unique and change is constant, companies must continually reassess their needs.
Technology hype and competitive pressures tend to frustrate strategic thinking, however. Instead of defining goals and identifying problems that need to be solved up front, organizations sometimes acquire technology or talent without a plan, which tends to negatively affect ROI.
"You need to have a really well-defined business case beforehand," said Jonathan Foley, VP of science at a recruiting software provider Gild, in an interview. "Companies are building out data science teams before they need them, before they understand what data science is and what is going to be the desired effect on the business. It's a me-too phenomenon where it's seen as something that can have a competitive advantage. But unless the leadership really understands the expected outcome of having data science and machine learning, it just becomes a difficult task. You don't know who to hire and you don't know how to manage the team once you have it."
[Are your data science efforts a success? Find out. Read 8 Ways You're Failing At Data Science.]
Data team structures vary significantly from company to company, which further complicates the issue. They may consist of internal and external resources, faithful employees, and new hires, or a combination of those and other things. Some organizations have multiple data teams, while others have comparatively modest resources.
There are centralized data teams, distributed data teams, and hybrids of the two. Despite the differences, there are some common best practices that can help just about any organization get more from its investments, regardless of size, industry, or budget. Here are a few things to consider.
**Elite 100 2016: DEADLINE EXTENDED TO JAN. 15, 2016** There's still time to be a part of the prestigious InformationWeek Elite 100! Submit your company's application by Jan. 15, 2016. You'll find instructions and a submission form here: InformationWeek's Elite 100 2016.
The race is on to use data more intelligently than your competitors do. As companies scramble to keep up, they often acquire technology and people without a game plan. Doing so can lead to costly mistakes.
"We strongly believe that the analytics are slave to the master of impact, so [it's important] to have a use-case-based approach: 'This is what I want to do and if I do that, I can generate these results and measure these results. Any technology that enables me to get those results faster or better is something we want to prove.'
Not every organization needs a data scientist or Hadoop. Nevertheless, companies are racing to fill positions based on the most popular keywords, rather than on the actual needs of the organization.
Technologies, companies, markets, industries, and business models are constantly changing. Often organizations are limited by their technology choices and the way they're utilizing data-related skill sets.
-
About the Author(s)
You May Also Like