Tips for Creating a Winning Data Scientist Team
Finding the right mix of support to do more with your data is no easy task. Data scientists remain in high-demand, and fetch top dollar. Here are some tips on how to assemble a winning team.
So much data, so little time
Organizations continue to struggle with how to get more out of their data. “It’s not a new challenge, but the problem is only exacerbated as more data is exchanged and created at petabyte scale,” confirms Dermot O’Connor, cofounder and vice president at Boxever. “The proliferation of data and the pressure for organizations to turn data into business value has increased demand for data science professionals.” Approximately 10 percent of the workforce at Boxever is data scientists, and O’Connor shared his views on how to best assemble a data science team.
Seeking the ‘total package’
“When a company seeks to hire a data scientist, it's typically seeking someone with skills in advanced programming and statistical analysis, along with expertise in a particular industry segment,” O’Connor explains. “The need is great, and the skills gap is widening: A study by McKinsey predicts that ‘by 2018, the U.S. alone may face a 50 percent to 60 percent gap between supply and requisite demand of deep analytic talent.’ Good data scientists are often referred to as ‘unicorns’ because it is so rare to find professionals who possess all the right skills to meet today’s requirements.”
Still the top job in America
“As the ‘top job in America in 2016,’ data scientists don’t come cheap,” O'Connor confirms. “How can today’s organizations harness the brains behind data science to get the most out of their investment, whether in talent or technology? Here are some things to consider when building your data science team…”
Data science is a team sport
“There are many facets to creating successful data science teams in a practical, operational sense,” O’Connor says. “It’s rare to hire just one or two on staff, so remember that for data scientists as much as any other role, strength comes in numbers.”
Outsource to innovate
“If you do the math, a team of seasoned data scientists – let’s say only five – will cost you well over $1 million annually in fixed costs,” O’Connor notes. “And like many in IT functions, they’re likely to be pulled in many directions. Having a dedicated resource to optimize your systems with networks getting increasingly smarter with every interaction via machine learning is one way to ensure that projects are efficient while blending technology platform costs with the costs for data science talent that drives them.”
Balance functional and strategic tasks
“Part of the reason data scientists are so in demand is because they have concrete skills in predictive analytics that others – in IT and business roles – lack,” O’Connor explains. “That being said, you’ll need sufficient talent and resources to both write and maintain software and algorithms while also gathering insights from internal teams and customers to customize and optimize the logic behind them.”
Set data scientists up for success with the right data management systems
“High volume, omni-channel systems are very complex – and time consuming – to manage,” says O’Connor. “Having a hub where data at the individual customer level is aggregated helps set the foundation for data scientists to really shine. Finding ways to automate processes so that the right data is available on demand will make any data scientist’s life easier and will make more possible under their strategic guidance.”
Expect to ‘see inside the black box’ of AI
“A data scientist should be tasked with explaining the process of machine learning and artificial intelligence in layman’s terms to bring in others into their realm throughout the enterprise,” O’Connor explains. “This is essential for gathering insights that make predictions stronger and actions more focused by design. And as marketers take on greater oversight of data, it’s important that CMOs and other decision-makers find complementary talent and technology to help them see the big picture to explore all that’s possible with their data.”