To assemble a high-functioning data science team, you need to invest some serious time, effort, and money to create a place where great data scientists can do great work. But how can you retain that hard-won talent for the long run? The traditional answer was to promote the best performers up the management career ladder. But have you considered that not all data scientists are interested in becoming managers?

As part of the release of our Winning the War for Data Science Talent report, I’m digging into the tactics of building a world-class data science team. Today I’ll focus on this recommendation from the report:

Create a non-management career path for your data scientists. Advertise it in your job descriptions.

You can sketch out career-advancement opportunities starting from your first interview with prospective hires. To learn more, listen to my recent discussion with Alex Cunliffe, one of our rockstar data scientists at Prolego.

‍

Management opportunities lack luster

The data science career study revealed what many of us expected: many data scientists aren’t enthusiastic about becoming a manager. The prospect excited fewer than 50% of data scientists in our study. But unfortunately management is the default career path for them.

‍

Management is hard. It involves a range of skills that can take years to learn. And of course, not everyone learns them. We have all seen the horrific consequences of bad managers. To serve the best interests of the data scientist, the team, and the company, avoid offering management as the only career growth option.

Career experiments

In our conversation, Alex raises a great insight from her own experience: many data scientists want the opportunity to experiment with career options before choosing one. Some might already know that an individual-contributor track or a management track is ideal for them. But many others will want to first test it.

Happiest work days

Alex says she’s happiest when she can code all day long. My best days are the opposite. Although I enjoy programming, I’m much happier when I can make someone else successful. It’s relatively easy to be decent at both programming and helping other people succeed. But it’s almost impossible to be truly great at both because of the required level of investment.

Three growth directions

As you chart out an alternative to the management career path, consider the skills your data scientists will need to develop to continue growing in their profession. Data scientists who pursue a non-management career path should invest in growth along three dimensions:

Domain depth. The best data scientists develop a deep understanding of data science. Lots of people can download and train a model from scikit-learn. Far fewer have the mathematical and theoretical knowledge to build a new solution from scratch.
Engineering breadth. Successful data scientists need the support of data engineers, MLOps engineers, product managers, data visualization teams, and many other professionals. Learning more about these domains will make a data scientist a more effective team contributor.‍
Team skills. Data science is a social skill. Effective data scientists must be proficient at explaining their work to leaders, engineers, customers, and other stakeholders.

Get a copy of Prolego’s data science career path

Now that you understand why you need to offer a non-management career path, you might want to see how to create an alternative path. As a bonus, our data science report includes an example of the career map and skill descriptions that Prolego offers to our data science team. You can download a free version of the report and our career map.