Big Data, Bigger Job

Everybody seems excited about the potential of huge datasets to turbo-charge their businesses. But do companies have the talent they’ll need?
Josh HyattApril 3, 2014

At Kevin Knapp’s most recent graduate school networking event, the chatter focused on “an issue that was close to the hearts of many of us,” as he puts it. No, he’s not talking about which of their classmates has gotten hitched most frequently or which fellow alum has funneled the fattest donations to the University of Chicago Booth School of Business, where Knapp earned his MBA.

Instead, the group exceeded its allotted time swapping ideas about, of all things, Big Data. Specifically, Knapp and his fellow alumni — the event was a regional gathering of about 50 Booth graduates who were finance executives — shared their concerns that the Big Data phenomenon “has emerged so quickly that there’s not as much talent around as companies need to get maximum value out of all the data.” Adds Knapp, who serves as CFO at CareerBuilder, the online recruiter and provider of human-capital solutions: “For CFOs, the constraining factor was related to finding people who could help them organize, set up and maintain Big Data systems. It was comforting to discover that this was a common issue many were struggling with.”

Drive Business Strategy and Growth

Drive Business Strategy and Growth

Learn how NetSuite Financial Management allows you to quickly and easily model what-if scenarios and generate reports.

13Sept_T_TS_TopArt big dataIndeed, Big Data is much more than the topic du jour. Beyond serving as an irresistibly compelling topic for conferences, blog posts and media spreads, Big Data is on its way to becoming a key differentiator for businesses in almost every industry, reshaping their strategy by enabling them to innovate and identify new opportunities. “It’s not a fad because it forms the basis for providing a real competitive edge,” says Gary Cokins, founder of Analytics-Based Performance Management, an advisory firm in Cary, N.C. “Using analytics to gain insights and foresights will give a company a lasting competitive advantage in the form of understanding how to make better, more accurate and quicker decisions.”

Those decisions, however, will still require human brainpower. While automated algorithms could conceivably overshadow the human element in decisions where speed is paramount — think of an insurance company, for example, which needs to know which claims require further investigation — the raw numbers aren’t going to speak for themselves. Gargantuan sets of data require interpreters who can program algorithms and prowl for patterns. In the past, such data scientists queried databases to find out if a certain hypothesis was true; now they can turn to the data to figure out the best hypothesis. “Today, the best leaders are not the ones with the best answers,” says Cokins, who previously served as enterprise performance management specialist at SAS, the vendor of business-analytics software. “The best leaders will be the ones with the best questions.”

Now those leaders are questioning how and where they will find the talent that can help them turn Big Data into Big Knowledge. “Data for its own sake isn’t going to get you anywhere,” says Kip Kelly, director of marketing and business development for executive development at the University of North Carolina’s Kenan-Flagler Business School. “Analysts who understand big data will be in high demand and ripe for poaching. Recruiting and retaining employees with these critical talents will become a challenge. Many organizations are worried about it, but they aren’t fully in the grips of it. They don’t know what they don’t know.” As consultant James Taylor, CEO of Decision Management Solutions, succinctly assesses the situation: “Most companies have neither the analytic talent they need now nor the Big Data talent they will ultimately need.”

CFOs may detect the early rumblings of an oncoming talent crunch. Chances are they’ve been visited by folks from different functions, from marketing to operations, appealing for the funds they need to hire or train analysts who can manage massive quantities of data — identifying reliable sources, scrubbing it, fitting it into a platform and applying complex algorithms to it. Accustomed to metrics and analytics as they are, finance executives need to know enough about Big Data’s challenges to serve as effective traffic cops, prioritizing the opportunities and communicating them with the IT function so as to keep that department from becoming paralyzed by all the pressure.

The Big Data mindset, after all — the notion that data is vital, not static — is sweeping not just through various business functions but also through entire industries. “Everyone is jumping on the Big Data bandwagon,” says Azam Foda, CFO of Univa, which makes data-center optimization software. “Everyone is asking, ‘Is it old wine in a new bottle, or is it something new?’ However, I’m not sure there are a lot of people who understand the heuristics and technology behind it.” Even those who think they comprehend Big Data may be intimidated by its potential power. As well they should be, perhaps. In the business realm, it’s being hailed as a tool as transformative as, say, electricity or antibiotics. Dating back to 2009, Google demonstrated how it could analyze patterns in search-engine queries to track flu outbreaks faster than the Centers for Disease Control. And while it’s true that Big Data analytics can help companies reduce costs, speed up time-to-market, boost productivity and increase profits, huge datasets alone don’t guarantee value.

The quantity of the data, in fact, may be a distraction. Many a Big Data discussion is dominated by conversations where participants “aren’t necessarily talking about the data-driven business problems they have, or what need, pain or problem [they] are trying to solve,” says Jill Dyché, vice president of best practices at SAS. “In conversation, it’s easy to get captivated with all of the emerging technologies around Big Data.”

Why Big Data is a Huge Deal

Just how big is Big Data? It’s fashionable to answer such queries by comparing Big Data’s size to an already inconceivably huge repository of information — the Ancient Library of Alexandria, the existing Library of Congress — and then recite just how much larger, exponentially, Big Data is. Another option is to blurt out a figure that is convincingly precise, if not wildly illuminating: at last count, there existed 1,200 exabytes of information (or 1 billion gigabytes), only 2 percent of which is non-digital.

What’s driving the Big Data movement is the fact that so much data, in so many forms, has been digitized. Even as text, numerical and transactional data continues to grow, information also piles up in the form of audio, images and video. It’s estimated, for example, that 400 million photos are uploaded to Facebook every day. Connectivity has given data-gatherers access to data from unusual sources, such as cheap sensors embedded in the physical world.

The ability to systematically and swiftly sift through such data to pick out patterns in, say, consumer sentiment about certain brands, is destined to redesign decision making and open up new opportunities. Consider that credit-score giant FICO has drawn from its large datasets a correlation between people’s credit scores and the likelihood that they will follow directions in taking prescription medication. The outcome: a new FICO Medication Adherence Score, which claims it can predict (with 70 percent accuracy) which patients are at highest risk for using prescription medications incorrectly. Such information is valuable to health-care companies, which can use it to send e-mail reminders to the most susceptible patients.

Across all industries, unlocking data will not only help determine a company’s competitive position, but will also be used to improve overall corporate performance. (Some sectors, such as those in the financial industry, will find it easier to capture value than others, having already made significant investments in IT.) Transparency, for instance, can drive efficiencies within operations, where supply-chain managers will gain better access to more of the data they need, both within their own corporate walls and inside those of their partners populating the supply chain. Data which exposes disparities in, say, the efficiency of different call centers can drive better performance from those who aren’t content to dwell in the bottom quartile. The ability to consult real-time results enables companies to engage in active experimentation, making better management decisions about, for example, what drives web page conversion rates. Furthermore, customers can be divided into micro-segments, receiving real-time offers based on their current location.

“We’re not only extracting and analyzing data, but we’re using it as well,” says Dyché. “That requires finding people with the brainpower and skills who can run analytics and communicate the results. These are people who can put what they’ve found into heat maps, bar charts and data-visualization tools — not just rows and columns. There’s a shortage of people who have the skill sets that businesses need to enlist in this new era of data scientists.”

To avoid being buried in an avalanche of data, then, companies may very well have to grow their own Big Data–savvy scientists.

Breeding a Better Geek

Maybe a dozen years ago, companies that needed statisticians — banks, for example — kept them comfortably stored in a dark room, trolling through data using brute-force tools. The sole interruption? “People would bring them water from time to time,” says Dyché. Cokin similarly describes the archetype as “geeks with PhDs who stayed in their cubicles.”

But with the invasion of Big Data, the rules of engagement have changed. There’s likely to be much more interaction between the data analysts and the rest of the business. At some companies, the data-crunchers will be embedded within certain functions, such as finance or marketing. At others, the brainpower will be clustered, with a pool of conversational quants available as a service to the company. Either way “these people aren’t going to be behind closed doors anymore,” says Dyché. “They’ll be waltzing around from user to user and from department to department, saying: ‘Look what I found!’”

The analyst’s interaction will no longer start with a manager filling out a form, petitioning for specific data. Now there’s much more likely to be face-to-face meetings, with managers sharing their objectives and goals. With access to better data, marketing executives may be compensated in part based on how effectively they act on that data — the return they’ve achieved on marketing investment. Data scientists may very well share in the same bonus plan. To reach such shared goals, the analysts will have to speak the same language as other executives.

The need to have their Big Data analysts understand their businesses has led big companies to, in effect, rebrand their statisticians. In early 2013, Dyché conducted interviews with 20 big companies to find out how Big Data fit into their overall data and analytics environment; she co-authored a study, Big Data in Big Companies, with Thomas H. Davenport, a visiting professor at Harvard Business School who has co-authored or edited four books about analytics. What they found, Dyché says, was that “executives were actually recruiting existing statisticians and business analysts into the data scientist role. The executives realize that incumbent staff already possesses the necessary relationships, and that they’re fluent in the business vocabulary. Recruiting internally enables them to hit the Big Data ground running.”

Many companies, however, are looking to universities to manufacture more data scientists; large vendors of advanced analytics tools have provided their software to these schools. Many universities have added online certificates and master’s of science degrees in areas like Big Data and Data Science. “With all the buzz about the shortage of data scientists, universities are jumping on it fast,” says Cokins. “But there will be a lead time.” Furthermore, adds Kelly, “Colleges and universities can crank out graduates, but they’ll really have to understand the business to use the analytics to drive the business. It takes a while to understand how the company makes money, and where you can leverage Big Data to drive the organization.”

Data scientists, of course, aren’t the only ones who will need to understand the possibilities inherent in Big Data. To a certain extent, analytics is soon to become a fundamental competency that everyone in a leadership position will need to have. “General managers need to know where the data lives, both internally and externally, and how they can ask questions of it,” says Kelly. “They’ll need to think more broadly about what they want to do with the business, and what information they need to figure that out. There is tons of data out there, from nontraditional sources, which is empowering. It should be driving your strategy.”’s Knapp says that “anybody who has preliminary experience with Big Data has seen its power, and wants to ramp up quickly to get there.”

In a recent CareerBuilder survey of more than 300 professionals at the IT executive, director and management levels, 88 percent of companies with Big Data–related plans report that they are struggling to find talent with the skills necessary to act on those initiatives. How is the talent shortage affecting their organization so far? Thirty-nine percent of participants reported difficulty making data-driven decisions, 36 percent said they are falling behind on projects, and 35 percent reported a mismatch between the skill sets of people leading Big Data initiatives and the skill sets needed to fulfill those missions.

Where will they come up with the talent to drive such efforts? CareerBuilder estimates that for every member of the “data engineer” category on its Supply & Demand portal, there are three to four open positions.

Bringing on candidates who don’t have relevant experience — but are willing to be retrained — can be a workable option. (See “Who Wants to be a Big Data Scientist?” at the end of this article.) “Companies are stepping in and sending people with non-IT backgrounds to get additional training,” Knapp says. “I don’t see CFOs saying that Big Data does not have a positive return on investment, so I think many would be willing to make that investment.”

No matter what they do, CFOs aren’t going to stop fretting about the Big Data talent gap anytime soon. “It’s something I’m focusing my attention on,” says Knapp. And come next spring, when his next alumni assembly comes around, he fully expects to take heart from hearing the continuing travails of his peers. “Companies will still be struggling to find the right people. Big Data is so new, and there’s no book on how to handle it. But we still have a lot to learn from talking to each other.”

Who Wants to Be a Data Scientist?

In early 2012, CareerBuilder embarked on an experiment: seeking to bridge the skills gap in IT, it created a six-month program to provide training for 10 participants. All of the class members were among the long-term unemployed, and most of them had no technical expertise. “These were people who had been in mortgage banking, or working mothers who had been at home for a number of years,” says Kevin Knapp, CareerBuilder’s CFO.

Serving as paid interns, the group was guided by company mentors, who worked with them on real-world business-intelligence projects. According to Knapp, several participants received job offers before the six-month course was finished; of those who completed the program, seven landed IT jobs outside of CareerBuilder, while one returned to school to earn a technology degree. “I was surprised at how many of these people found jobs in Big Data right away,” says Knapp, who reports that the company is now assembling its third group of IT interns.

“This didn’t cost us a lot. It cost our people some time,” says Knapp. “And it shows that we don’t have to wait for the education system to fix this.”