CIOs must encourage data scientists to solve real business problems, not just play with data.
As more organizations hire data scientists–especially for predictive analytics projects–IT leaders are discovering that managing people who can turn data into ideas for business actions takes a deft touch. The sharp analytical skills key to the role can sometimes get in the way of answering big-picture corporate questions.
“I’m coaching them to make sure they’re aligned with the company, but I’m not prescribing methodology,” says Anne Robinson, director of supply chain strategy and analytics at Verizon Wireless. “Because if you want a high return on your analytical investment, allow them the freedom to explore.”
Robinson, who is also president of the Institute for Operations Research and the Management Sciences, a professional association, says good teams incorporate a mix of academic skills and applied experiences. Personal characteristics, such as the ability to make connections and express ideas well, are important in the corporate setting, managers say.
And it’s increasingly important for CIOs to be able manage analytics teams well–an Accenture study last year found that analytics is “moving from a secondary role in business to the core of many key decisions and processes.” For example, analytics may be used to predict customer behavior or prescribe changes that make the supply chain more efficient.
Working in the Real World
Managers should guide their data scientists to interpret data, not just crunch it, says Betsy Page Sigman, a professor at Georgetown University. “Some data scientists are so fascinated by data they lose the forest for the trees,” she says. Focus them on bigger corporate goals so they can make predictions in a business context.
Andrew Jennings, chief analytics officer at FICO, a $676 million financial services and credit score company, says statistical skills are hardly enough. He wants people who can both program and see how analytics can be used to shape business strategy. “It’s absolutely critical to understand the problem you’re trying to solve,” he says.
For example, if his team is working on a predictive analytics problem such as improving fraud detection at the point of sale, it needs to analyze the data and factor in real-world business conditions, such as the need for speed and no false positives in the final product.
Finding all those skills in one person is tough, so Jennings looks at the team as a whole. Team members fill roles that use their strengths: a data scientist with communication skills, for instance, will work with business folks.
Other traits are also important. Lon O’Donnell, manager of professional services at International Game Technology, tries to foster inquisitiveness in the data scientists at the $2.2 billion gaming systems company.
“I need someone who makes sense of the data instead of just aggregating it,” O’Donnell says. He wants people willing to learn the gaming industry.
In turn, to keep high-performers happy, he stockpiles less urgent projects to provide challenges during slow work times. “You have to always engage their minds,” O’Donnell says.
By: Mary K. Pratt. Originally published at news.idg.no
Read more about big data in CIO’s Big Data Drilldown.