It's a rare organization that would argue against analyzing past performance. But using data as a kind of Magic 8 Ball? That makes for a harder sell.
Or so you would think. These days, organizations are increasingly interested in data mining, predictive analytics and forecasting to better prepare for what's to come. But integrating the advanced techniques -- and finding talent with the analytics skills to take on the task -- can be difficult, as can proving the data's validity to executives.
"You have to get some early wins on smaller projects that can show you need more than Excel software," said Tim Rey, director of analytics at The Dow Chemical Company, headquartered in Midland, Mich.
That was one suggestion Rey had for those attending his session on accurately predicting future outcomes at SAS Institute Inc.'s recent Premier Business Leadership Series. He was joined by Jeremy TerBush, vice president of analytics for Wyndham Exchange & Rentals, part of Wyndham Worldwide, based in Parsippany, N.J. Together, they shared insights on how to build an analytics team and how to find new talent.
Finding quantitative skills
Both Dow and Wyndham have a roughly 30-member business intelligence (BI) and analytics team. Each company divvies up its team along lines of functionality: Modelers model, BI practitioners practice BI and so on. But finding talent with advanced analytics skills can be a challenge, especially as the data -- and data needs -- continue to grow across the business landscape.
"It's definitely going to be hard in the next two or three years," Rey said. "There is very much a shortage of people in this space."
According to a frequently quoted projection from a 2011 McKinsey & Company report on "big data," the U.S. could face a skills shortage of 140,000 to 190,000 people with the necessary analytics skills by 2018.
To offset the challenge, Rey suggested searching for talent from within the organization first, as Dow Chemical did when it began building its analytics program. Early on, he said, the team looked at the organization's Six Sigma program, developing data mining training to bring in interested employees and get them up to speed.
"[The idea was to] see if they aspire to do something more quantitative and then pull them in," said Rey. "You can train them with some of the specifics."
More likely than not, businesses won't be able to fill their data practitioner needs with internal talent alone. Rey also encouraged forming partnerships with local universities to build "a feeder stock." Here, he and TerBush cautioned businesses against restricting searches to candidates with a certain set of skills; instead, cast a wider net, looking for potential talent from disciplines such as engineering, physics and even economics.
"We have generally taken the approach of hiring people who are pretty green … and we build up their analytical capabilities," TerBush said. "We've been pretty successful with that approach."
Building analytics bridges
Beefing up the company's analytics program doesn't begin and end with quantitative talent. The team also needs to prove its worth to the company. To do that, Rey and TerBush recommended finding projects the team can quickly and assuredly add value to.
"As soon as you get a couple of [wins], your customer base will start to grow," said TerBush. "Pretty soon, you will not have enough people [to handle the demand], which gives you justification to expand the team."
Rey and TerBush also suggested waiting to invest in big software purchases, pushing the process forward even without perfect data and obtaining executive-level support.
"It helps a lot," said TerBush. "And it helps to break down a lot of doors that might be closed to you initially."
That last piece of advice, though, leaves analytics teams at the mercy of chance. TerBush's CEO, for example, is data driven, which has inspired a greater interest in analytics across the company, he said. But for others, data -- which usually contains a margin of error -- can be met with reluctance and even mistrust by executives as well as the line of business.
Rather than generate an adversarial relationship, Rey suggested proposing analytics as a (not the) decision-making tool. Analytics can act as a comparison for gut instinct and intuition, all while helping to create standards and mathematically test assumptions.
"One of the values of doing analytics is actually reducing risk. It's confirming what you know so that you have a comfort level," he said. "And I think that's more of a positive relationship you can have with analytics."
A shift in perspective is one way to help change the culture of a company. Another is investing in communication between IT and the line of business, though, according to TerBush, filling that role is becoming even more difficult than finding quantitative talent.
"We find it's harder to find … people who can really speak to the business and who can translate it back to the technical side," said TerBush.
A techie, business sort of hybrid
As Wyndham grows and acquires less analytically savvy companies along the way, TerBush tries to identify what he calls engagement managers, people who can understand both the technical and business sides of an organization and move seamlessly between them.
Rey referred to this role as the business analyst, which he believes will become increasingly important for businesses, just as analytic skills have become.
"They are not the Ph.D. or masters in statistics or machine learning," he said. "They'll be the translators between core competency and business."
Otherwise, Rey said, look for innovators willing to take a chance and build that portfolio of success stories and testimonials. And don't shy away from marketing the talents of the department.
"Get out the white shoe polish," he said. "You need white shoes and a white belt and [you need] to start selling."
Research shows predictive analytics programs require ongoing training
Find out why storytelling is a necessary analytics skill
Learn how 'big data' is helping to breed a skills shortage