Every week an AI leader at a Fortune 1000 company asks us the same question:

We’re having a hard time attracting and retaining data scientists. Do you have any suggestions?

After many of these conversations and years of working closely within our clients’ environment, I see consistent root causes for the talent challenge. In this post I’ll describe many of the most common problems.

You haven’t accepted that the talent is in control

In the pre-AI era, you were in control. The talent supply outstripped demand, and you had choices. Your new strategy must start by accepting that the talent is now in control. Opportunities for qualified data scientists exceed the number of people who can fill them. This is a permanent market shift, and conditions will continue to worsen for employers.

Your company is in denial 

Although you might have accepted the reality of the talent marketplace shift, your company is in denial. The current crisis is different from the talent shortage of the dot-com era, and it won’t go back to “normal” anytime soon.

Your leadership doesn’t want to admit that your company has the wrong technology workforce for the AI era. HR doesn’t want to change its recruiting processes or job description templates to sell the new data science opportunity. Recruiters are trying to convince you that the data scientist whose only credentials are online certifications is qualified to solve your hardest problems. Leadership is resisting your requests to hire remote data scientists.

You’re starting to realize that you can’t succeed unless you can wake them up. (This moment of realization is when people like you reach out for help. It’s why companies hire us—to help make the case to their leadership that the new landscape requires a new game plan.)

You don’t have an AI mandate

As our data science workforce study revealed, many data scientists leave their jobs because their company isn’t serious about becoming AI-driven. Yes, many companies are aggressively recruiting data scientists and paying top-of-the-market salaries. But they aren’t investing in the data, infrastructure, or organizational changes that will allow a data scientist to succeed. Unless your company has a clear mandate for change, the data scientists will soon realize their efforts won’t make an impact. So they’ll leave.

You haven’t defined the data scientists’ role

The job title “data scientist” is overloaded. It can mean everything from a business analyst to a software engineer, depending on the company. 

But the skills required to do basic data cleansing and modeling are much different from the skills needed to reinvent your business with a new methodology. Unfortunately too many companies fail to understand this distinction until they realize they’ve hired the wrong people.

Your data scientists don’t do data science

Some companies recruit data science talent by offering big salaries, lofty titles, and the promise of interesting work. And after they onboard the new talent, they break their promise by tasking their new data scientists with systems integration projects, data cleansing, app development, dashboards, or any number of projects unrelated to data science.

Top performers place a high value on learning and keeping pace with the industry. They won’t stick around for a title and salary at the risk of letting their skills atrophy.

Your hiring process is tedious

Your attention to the recruiting process says a lot more about your culture than the job description does. Too many companies drag out the interview process or ask candidates to work on unrelated take-home coding assignments or problems. Unnecessary obstacles will steer qualified candidates to competitors who are optimizing their process.

You’re ignoring your company’s unique opportunities

Why are you comparing yourself to Facebook and Google? Lots of talented people avoid job opportunities there for many reasons, including the ethics of their advertising business. If your company is in a legacy industry, do you emphasize how potential hires can use the job to affect society for good? Do you advertise the opportunity to transform an entire industry? Are your projects aligned with socially responsible goals, such as improving health or reducing our carbon footprint?

Your hiring managers misunderstand data science

If your  hiring manager comes from a traditional software background, they might not understand cutting-edge data scientists. They might not even know what to ask a candidate. During an interview, if the hiring manager conveys the idea that the purpose for the job is to produce a model that the team can integrate into their existing framework, experienced data scientists will see warning signs that the job is a dead end.

Your poor performers push talent away

One terrible consequence of the “great resignation” is that your best people left. The remaining team members are likely your lowest performers. After all, the first employees to seek other opportunities are those who have choices. 

As the overall quality of your team falls, the remaining good and average performers are forced to take on more responsibilities or deal with jerks. Soon these people leave, and the problem compounds itself. In this kind of environment, a good data science team can completely dissolve within a month.

Your lousy bosses are driving talent out the door

Take 30 seconds and think about the worst boss you’ve ever worked for. Now ask yourself how long you would have kept your position if you could have landed a better job in a week simply by responding to one of the many recruiters chasing you on LinkedIn. 

If your team keeps losing talent, take a hard look at the boss. Are they hard to work with? Do they lack vision? Do they give the team conflicting and inconsistent direction? If so, ask yourself how much their behavior is costing you.

You’re spending too little time on recruiting and retaining talent

I’m often shocked at how little time companies actually invest in their people. If I ask 100 AI leaders (or actually any leader) about their biggest challenges, 99 of them will say it’s finding and keeping the right people. If I then ask them how many days per week they spend on talent, they inevitably start talking about how busy they are with other obligations. And in fairness, they are overwhelmed. But they won’t solve their biggest problems without investing time in their human resources. We have specific tips in our AI workforce report.

You’re not tracking or acting on talent metrics

Do you know what your data scientists want from their job, what they fear, or why they leave? Have you surveyed the people who left or who turned down opportunities? Do you track your recruiting process by using techniques from digital marketing and sales automation? Are you slower or faster than your competitors? Are you improving? 

If you don’t know the answers to these questions, you need to start asking them and tracking what you find. If you do know the answers to these questions, it’s time to act on that knowledge.


To attract great data scientists to your company and retain them on your team:

  • Enter negotiations with the understanding that they have formidable bargaining power.
  • Get your company’s buy-in.
  • Formulate a clear and compelling AI mandate.
  • Clearly define job roles.
  • Provide fulfilling work.
  • Streamline your hiring process.
  • Promote your company’s selling points.
  • Make sure your hiring managers understand the breadth and importance of the job role.
  • Improve the team’s talent environment.
  • Take a hard look at your team managers.
  • Spend time with your data science team.
  • Track and act on talent metrics.

Want to understand more? Check out our data science workforce study. We surveyed hundreds of data scientists to learn the secrets to attracting and retaining this emerging workforce.

Kevin Dewalt
Chief Executive Officer & co-founder

More Ideas

AI Abundance:

Why you have only five years to prepare for the inevitable business extinction event.