Meet the Leader of a Data-Driven Work Culture: The Data Science Manager Meet the Leader of a Data-Driven Work Culture: The Data Science Manager
What does it take to build a data-driven culture at a company? Surely, you’d expect hiring a few people with technical... Meet the Leader of a Data-Driven Work Culture: The Data Science Manager

What does it take to build a data-driven culture at a company? Surely, you’d expect hiring a few people with technical skills would make a monumental difference.

But Google Chief Decision Scientist Cassie Kozyrkov says the true key to building a data-driven culture is hiring a data science manager — an often-neglected role that makes or breaks these teams.  

Kate Strachnyi, a manager at Deloitte and data visualization specialist, describes what the role entails and how data scientists can work toward a managerial role.

[Related Article: A Data Science Guide for Managers]

What is the job?

Of around 15,600 data science jobs Indeed.com displayed on Nov. 8, around 19 percent included the word “manager” in the title. They were at places including Google, BuzzFeed, Amazon, Walmart, Big Fish Games, and Facebook.

Data science jobs that did and didn’t include a managerial role had many of the same responsibilities. Both included roles extracting, cleaning, and analyzing data. Both required knowledge of coding languages including Python, R, and SQL. Many job postings in both mentioned machine learning and artificial intelligence to develop predictive models.

Nearly all positions indicated the data scientist must use data to find and communicate business solutions to other, non-technical teams and build potential strategic roadmaps for the company.

But many manager postings didn’t emphasize these technical skills as much. Instead, managerial tasks focused on planning alongside senior leaders, fostering innovation, overseeing experimentation, and leading a group of data scientists.

Strachnyi said in her role at Deloitte, she oversees a team of two. The team is in charge of internal-level reporting on finance and talent. The trio uses data from several sources to create dashboards and give executives insights, like where they’re not generating enough revenue.

Before she became a manager, Strachnyi said she managed projects instead of people. Now she keeps tasks on budget and on time, and keeps her team aware of high-priority projects for the week. Rather than consulting with Deloitte’s clients, she consults with executives.

“I’m thinking a lot more about, ‘What could they ask of us in next week or so?’ versus being reactive in my prior role, where I would just do whatever was asked of me,” she said.

What extra skills do they need?

Each company has different management needs. Walmart, for instance, requests applicants have experience leading and mentoring a data science team for more than three years. BuzzFeed wants its data science manager to “practice radical candor” and give direct, actionable feedback to people they do and don’t manage.

Amazon asks that data science managers demonstrate excellent verbal and written communication skills. That’s because oftentimes, managers must present the data science team’s findings to company leadership to inform their decisions.

Indeed, Strachnyi said communication and presentation are the most significant soft skills a data science manager needs. Beyond the technical skills to perform the analysis, this kind of communication requires a deep business understanding of the company and industry.

“Being able to relay your analysis results is so important, especially if you can say it in language your audience can understand,” she said.

Otherwise, data science managers must be adept at any relevant technical skills so they can lead their teams by example. Most companies wanted their managers to have a master’s or doctorate degree in a computer science-related field. Some ask for a decade of experience.

How can you prove you’re ready to be a manager?

Less than three years after starting at Deloitte, Strachnyi decided to go for a promotion. She said company leaders don’t automatically distribute promotions in this industry. Instead, workers must tell their coaches that they are ready and start a conversation.

Strachnyi had to build a business case of how her work managing projects would translate into managing people. She said managing the scope of work on a project is comparable to managing people because she has to think of what she can delegate to her teammates. She aims to make them feel empowered, but also challenged by the work.

“There’s a fine line between scaring the person below you with too much work versus them being bored,” she said. “There’s a big personal factor where you have to spend time with your team to see what their goals are.”

One of the most crucial things a data scientist can do to show they’re ready is demonstrate business/industry understanding.

“You can know the technological stuff, but if you’re not sure why you’re doing the analysis and how that’s driving revenue or saving money, I don’t think you’ll be moved up in the ranks,” she said. “You have to understand how everything relates back to the business.”

She also read books about about management, leadership, and how to influence people. What’s vital, she said, is to continue to learn, grow, and stay up to date on trends and what’s happening in the world.

[Related Article: Managing Effective Data Science Teams]

Strachnyi was pregnant when she decided to go for a promotion, a time when she said many women might start to lean out of a career for a bit. But she said it was important to realize that she was ready to move up and seize that opportunity, and encourages others to do the same.

Strachnyi’s thinks her go-getter attitude and willingness to do whatever it took to get the job done earned her the position. How did you demonstrate your readiness to your superiors? Leave a comment or tweet us @ODSC so we can share it with other data scientists aspiring for higher roles.

Paxtyn Merten

Paxtyn Merten

Paxtyn is a student at Northeastern University studying journalism and data science.