Webinar: Why logical layers matter, and how to use them -Watch now

The Big Four Reasons Companies Struggle to Hire Data Talent

Image of author
Derek Steer, Co-founder & Chief Strategy Officer

July 11, 2018

4 minute read


Between the thousands of data scientists, analysts and others who use Mode every day, and the hundreds of companies that run their analytical operations on Mode, we hear from both sides of the data talent market. From these conversations, we've noticed some common reasons that companies struggle to hire data talent:

1. Companies don't know what they want.

Eric Mayefsky, Head of Data Science at Quora, has done a good job articulating how data scientists should decide what they're looking for in a role. He recommends candidates be particularly thoughtful about the type of role they are looking for-- what flavor of data science they want to practice. This process is equally if not more important for the companies doing the hiring. To borrow a phrase from Insight Data Science CEO Jake Klamka, there's a lot of “and, and, and” in data science job descriptions.

How to fix it: Organizations that want to hire data talent need to make it clear what goals the person is being hired to meet, and what resources are available to help meet them.

2. They don't have hard data problems.

Not only do they not have hard data problems, they don't know how to match talent to the problems they do have. This is a more specific (and very common) extension of the previous point. Many companies attempt to hire for a job they wish they had, when they really need to hire for a different job that must get done first. For example, a company might be trying to hire a machine learning specialist when their data pipeline hasn't even been built yet. There are many talented candidates out there who want to work with the latest technology or solve problems in really complex ways. But the reality is that most jobs don't require that, and hiring someone who wants to solve problems in unnecessarily complicated ways could be counterproductive.

How to fix it: Hiring managers need to communicate as much what the job isn't as what the job is, and filter out poor-fit candidates early with detailed job reqs and careful screening.

3. Hiring managers don't actually know how to evaluate data scientists.

All too often, technical skills are the only barometer used to screen people out in data science recruiting processes. Everyone knows that communication and problem-solving skills are what make for really great teammates, but nobody seems to know how to actually screen for those things. Organizations that neglect to formally examine the non-technical skills of the people they interview will end up making offers to candidates with deep technical knowledge and few other strengths.

How to fix it: Learn how to interview data scientists on all three key areas.

4. They don't value or understand the practice of data science.

And some simply don't care. This manifests in lots of ways that great candidates can sniff out quickly. Who does the role report to? Does that person think Analytics or Data Science is valuable? How do they see the data team contributing to the company as a whole? As an example, I very rarely see Analytics thrive when it reports through Finance, though Intercom's team is a notable exception. Really talented analysts don't want to be dashboard monkeys, and they will turn and run if they sense that their manager thinks of the role this way.

There are other signals here as well. How much budget is allocated to the team? Can they buy the tools that they will need to be really successful? Can they hire supporting roles like data engineers or product managers? Basically anything that communicates the notion of “we just want it to work” will send talented folks out the door and on to the many other places looking to hire them.

How to fix it: Ask yourself the questions above and think hard about whether you would want to take that job. If the answer is no, you may need to rethink the value you're placing on this work- otherwise, you may be left behind.

The good news is, none of these problems are fatal. They may be challenging, depending on your company's unique situation. But if an organization decides it's worth it, all these problems can be solved. And if an organization thinks their problems can't be solved...refer back to problem #4.

Growing your data team, or looking for your next gig? Post a job or find one on Mode's Data Jobs Board.

Related Articles

Get our weekly data newsletter

Work-related distractions for data enthusiasts.