A guide from the head of HR at a leading analytics company
Data scientists are in higher demand than ever before. According to the latest CrowdFlower survey, 79% of respondents reported a data scientist shortage in 2015. In 2016, that number grew to 83%. The race is on to find skilled people who can organize, structure, and make business sense out of Big Data sets. People with heavy STEM, analytics, and conceptual skills, and the attendant work-friendly personality traits (insatiable curiosity, ability to prioritize, and a healthy dose of skepticism, to name a few), can virtually write their own tickets.
Recruiting and retaining top-flight data scientists and other professionals in the data science field requires an understanding of their personalities and motivations above and beyond escalating salaries, free lunches, and other perks. Here’s what we’ve found resonates with this highly sought-after group.
Brand Affinity
Brand awareness and reputation cuts both ways among the data science crowd. Many want to work at Google, Amazon, or other giants known for pioneering work in artificial intelligence (AI) and machine learning. This is especially true for younger players who want to validate their bona fides and get a marquee credential as a career springboard. Tech giants like these typically have no problems finding talent.
Companies that don’t have immediate brand recognition need to flex some other muscles. We’ve found that many data professionals warm to the prospect of working for non-tech brands offering an opportunity to apply data science to commercial or societal challenges. Others prefer scrappy, data-heavy startups, where equity and recognition are options. These folks often see value in playing a bigger role in a smaller environment earlier in their career.
Peers
Data scientists want to be constantly learning from peers and finding potential mentors. The number, experience, and pedigrees of peers should be part of any recruitment story. Strong data science teams are persistent learning labs where ideas, techniques, and approaches are shared, debated, tested, and refined.
Present your data science bench and position a recruit in the environment in ways they can visualize or anticipate their role. Managers with data chops need to be closely involved with recruitment and offer a keen understanding of the math and the psychology of being a data scientist.
Tout Toolsets
Data scientists want to be on the cutting edge, using and maybe even inventing new tools. Be sure your IT infrastructure and equipment are near the latest generation with maximum capacity. If you’re simply doing serial regressions with outdated software, or using legacy databases and not working in Hadoop, don’t bother seeking out these individuals.
New recruits in particular want to learn from expert-level users of R, Python, deep neural networks, and trainable models. Of-the-moment software and analyticstechniques are very attractive to data scientists. Everyone is thinking ahead. Positioning your toolset as a career roadmap is critical for attracting people who are religious about adding skills to their personal toolbox.
Real-World Challenges
Everybody wants to work on something that matters. Data people have a Trekkie-like focus on going where no one has gone before. Working on big challenges for well-known companies and brands is a big draw. Solving complex business problems by finding something new or different in the data is extremely gratifying and highly motivating.
Effectively recruiting data scientists requires exposure to the problem or challenge being addressed. A corollary is explaining context: why the work matters, how it fits into a company’s business model, and how the work might advance the field of data science. On a day-to-day basis, the work of a data scientist can be tedious and unappreciated. So it’s important to share the big picture and show how their work will matter — even if it requires long hours of solitary number crunching.
Balance between data prep and problem solving
In most cases, an enormous amount of data ingestion, preparation, and normalization must take place before anyone can start the “fun” part of advanced data science. Data scientists want to know the ratio of rote or routine work to interesting predictive analysis they can expect to do.
The answer may depend on the available toolset, the department’s internal organization, the company’s management tactics, or a combination thereof. Recruiting top talent requires an attractive ratio of routine to breakthrough work assignments and a modest amount of self-directed work or choice of assignments.
Conclusion
Finding and landing top data professionals requires an understanding of their needs and a competitive commitment to super-serve them. Managers and companies must engage data recruits on their own terms and paint an enticing picture of the brand, the environment, and the challenges, as well as illustrate what to expect in terms of their workflow and achievements.
Tim Bridges is Senior Vice President, Head of Human Capital Management at Opera Solutions.
Danny Flamberg also contributed to this story.