Nov 02, 2012 (07:11 AM EDT)
Big Data Talent War: 7 Ways To Win
Read the Original Article at InformationWeek
The just-released InformationWeek 2012 State of IT Staffing Survey reveals that 40% of those who cite big data and analytics as a top hiring priority say they'll increase staffing in these areas by 11% or more during the next two years. At the same time, 53% of these companies say it will be hard to find big-data-savvy analytics experts. Respondents expect to try a mix of retraining of existing people, hiring of new employees and contracting of consultants and temporary employees to fill the gap.
Practitioners, vendors, and educators we spoke to for our Big Data IT Staffing report offer seven tips for finding the right talent.
First, have existing employees attend conferences, webinars and vendor-sponsored training classes that offer low-cost educational opportunities either locally or online. It's a great way to find out who's prepared to lead your big-data initiatives and what holes you need to fill.
Second, institute a liberal tuition-reimbursement program to cover a range of educational opportunities. That's what 65% of our respondents say they've done as an incentive for employees to retrain.
Our third tip is to rethink how you're taking advantage of existing talent. Many large and sophisticated companies have analytics experts on staff, but their work is often confined to areas such as research and development. Dow Chemical, for one, has successfully reassigned many of its PhD-caliber employees from R&D to work with business units on operational challenges such as optimizing supply chains, logistics, purchasing, and pricing.
Our fourth tip isn't for every company, but even comparatively small outfits such as Ancestry.com go to where the talent is available. The genealogy website has its headquarters in Provo, Utah, and a satellite office in San Francisco. General Electric opened an office in San Ramon, Calif., specifically to draw on tech talent in the San Francisco Bay area.
Analytics has been a hot topic for at least five years, so many colleges and universities are now turning out graduates from newly established degree programs. Our fifth tip is to tap into standout analytics programs including those at North Carolina State University, the University of Ottawa, Northwestern University, DePaul University, the University of Connecticut, Oklahoma State, Texas A&M, Texas Tech, California State University at Long Beach and the University of Alabama. Schools offering degree programs in the big-data-oriented discipline of machine learning include Carnegie Mellon, California Polytechnic State University in San Luis Obispo and the University of California at Berkeley.
Our sixth and seventh tips relate to the skills you should look for and the corporate culture you should cultivate as a would-be employer. The advice cuts both ways if you're a would-be big data analytics employee; read on to consider our detailed advice on what skills to build and the type of environment where you might feel at home.
Just how strong is the demand for big data analytics experts? To give you some idea, 40% of 108 respondents to our InformationWeek 2012 State of IT Staffing Survey who cited big data and analytics as one of the top two areas of staffing increase over the next year say personnel will increase by 11% or more during the next two years. And 18% say they'll increase staffing by more than 30%.
More than half (53%) of our survey respondents say required big data analytics skill sets will be hard to find. Another 23% say the salaries demanded by the people with such skills may be beyond their budgets. Only 17% were optimistic that they will easily fill the positions they have available.
There's no single prescription for filling available big data analytics positions. Among respondents intent on big data hiring, 33% say they'll "do a mix of retraining and hiring/contracting." The second-largest group, 28%, will "mostly retrain staff and hire/contract a few people." The third-most popular choice is to "hire or contract to fill needs," cited by 16%. The smallest group, 11%, will "retrain staff we already have."
Companies won't be able to fill the talent gap with recent graduates and people lured away from other companies. Fortunately, you won't have to beg existing employees to line up for training opportunities. If you've attended a Hadoop or NoSQL-related conference in the last year, you've undoubtedly seen the throngs of 20- and 30-something data geeks (and sometimes 40- and 50-something geeks) packing into the keynotes and seminars.
Big data events put on by show organizers and vendors alike are serving up a growing diet of educational content. And there are myriad online courses, webinars and certification programs available. Hadoop software distributors Cloudera, Hortonworks and MapR have all made training a big part of their business models, and most of the open source and commercial NoSQL database providers have done the same.
An overwhelming majority (65%) of our big data-focused survey respondents say their companies pay for training as an incentive. Another 39% say their companies offer performance incentives to those who update their skills. And 33% say raises and title increases are used as an incentive to improve big data analytics skills.
Large companies and sophisticated companies often have analytics experts on staff. They're usually found in the research and development or finance departments. But some companies are pushing these groups to share the expertise.
Dow Chemical wanted to get more predictive throughout its business, so in 2005 it kicked off an experiment in which two analytics experts from R&D were asked to help out in operational areas. These experts helped the purchasing department develop a freight and logistics cost model to analyze about $2.8 billion in annual truck, rail, ship and airfreight costs worldwide. A supply chain effort led to a model to analyze $4 billion in annual raw materials spending. Both efforts helped Dow save big bucks by accurately predicting costs and enabling procurement people to buy early or wait to buy on better terms, reducing cost by renegotiating contracts.
Early successes at Dow led to a corporate-wide initiative in 2010 through which it has shifted 10 of its Ph.D.-level analytics professionals to work full time with business units to develop predictive and statistical forecasts. Enhanced sales forecasts backed by advanced analytics have reduced forecasting errors. Business units now know by mid-month whether they'll meet monthly performance targets so they can adjust their strategies accordingly. Exchange rate and margin analyses have helped Dow make decisions about where to buy raw materials and how to determine pricing of finished products.
There's clearly strong demand for anybody with expertise in analytics or big data management. People with experience in both areas are "going for crazy prices," says Brian Courtney, general manager of industrial data intelligence at GE Intelligent Platforms (GEIP).
"We have a lot of analytics talent on staff already, but getting high-end analysts with big data experience is that much harder," says Courtney, who's part of a $200 million, 3,500-employee business within GE that offers big data analytics software and services to the industrial commercial software market (including customers both inside and outside of GE).
GEIP's specialty is big data analytics software and services, so part of Courtney's job is getting the word out to potential employees that GE is hiring. Proximity helps; GE has offices all over the world. But it opened a Software Center of Excellence in San Ramon, Calif., specifically to attract employees in the tech-talent-rich San Francisco Bay area. Even comparatively small companies are going to where they can find the talent. Genealogy website Ancestry.com has an office in San Francisco to draw on that talent pool while headquarters is in tech-endowed Provo, Utah.
When big technology waves come along, as we've seen in the last five years with big data analytics, it takes about 10 years to train the next generation on the new skills that are needed, according to Jim Spohrer, IBM's director of global university programs. "We're in one of those 10-year cycles right now where we're shifting to create the next generation of graduates," he says.
Thus a new wave of graduates is just starting to emerge and will reach a steady stream by 2018. Spohrer tracks IBM's work with various colleges and universities, including financial support, research grants, software and technology donations, and recruiting. Among the institutions on Spohrer's short list of top analytics schools are North Carolina State University (NCSU), the University of Ottawa, Northwestern University, DePaul University and the University of Connecticut. At SAS, Jerry Oglesby, senior director of global academic programs, also cites NCSU and Northwestern as leaders, but he also lists Louisiana State University, Oklahoma State, Texas A&M, Texas Tech, California State University at Long Beach and the University of Alabama.
Genealogy website Ancestry.com needs big data scientists who can work with Hadoop and write their own algorithms. To find employees who are hip to R statistical programming and MapReduce, the company is working with schools that have introduced course work or degree programs in machine learning. Scott Sorensen, senior VP of engineering, says institutions including Carnegie Mellon, California Polytechnic State University in San Luis Obispo and the University of California at Berkeley are among many that have stepped up their machine-learning programs.
The new analytics degree program at Louisiana State University (LSU) combines the computer science topics of data management and business intelligence with training in statistics, predictive analytics and operations research. There's also a focus on applied training in areas such as fraud detection, risk management, text mining and process improvement.
"Employers are looking for people who understand the business side but who also have an understanding of what IT and statistics can do to solve business problems," says LSU's Helmut Schneider, who helped set up LSU's analytics curriculum.
Attracting experienced employees from other companies is not just a matter of using social networks like LinkedIn and offering more money. Talented people want to know that they'll be working with the latest technologies, that they'll have access to training and that they'll be collaborating with like-minded employees. What's more, they want to work on important insights that help drive the business.
It's easy to see the potential culture clashes at a big data event near you. Hint: The attendees from banks, insurance companies, manufacturers and chemical companies aren't generally wearing hoodies. The stubble-and-T-shirt crowd typically hails from startups, media companies or Internet giants where foosball tables aren't just unused affectations purchased to make a company seem hip. Skills and cultures don't always work well together. Keep that in mind in your quest to hire or be hired.