Should Companies Worry About Analytics Skills Shortage?


With practically every new technology, there's an accompanying shortage of professionals equipped with the right skills. We've seen skills shortages with service-oriented architecture, and companies clamoring for folks with virtualization skills.


Now it appears they may soon be shaking the trees for statisticians, a position that Google's Hal Varian predicts will be the "next sexy job" in IT.


Companies would do well to consider a possible shortfall in statisticians before making any big investments in analytics software, advises Intelligent Enterprise Editor-in-Chief Doug Henschen. While vendors provide prebuilt models and off-the-shelf analytics applications, many companies find they need to customize to gain insights that are truly relevant to their business. He cites one, wholesale auto lender Dealer Services, which tried off-the-shelf analytics but is now building its own predictive model incorporating dealer traits and transaction parameters that have worked well for the company in the past.


These kinds of data-related jobs made a good showing in summer job listings. While July marked a record low for listed openings, according to the Labor Department's Job Openings and Labor Turnover survey, the Conference Board tracked an increase in jobs involving computer and mathematical sciences, from 398,000 in July to 406,800 in August. The Associated Press story specifically mentions IBM hiring or retraining people to work on data analysis projects, something I also referenced in my statistician post.


Interestingly, as Henschen points out, companies don't appear to be unduly concerned by the prospect of a shortage. Thirty-four percent of the 534 participants in a recent InformationWeek/Intelligent Enterprise survey said they "already have skilled analytics professionals on staff." Forty-eight percent said they expect to "train in-house BI experts and power users on analytic tools."


Henschen suggests paying close attention to users' comfort (or lack of comfort) with analytics software in any pilot projects and determining whether they can be trained to use analytics apps without lots of ongoing intervention from consultants. That point is echoed in a post I wrote last year. Responding to an earlier post of mine, reader Simon Mortimore warns that not performing a skills assessmentand providing training when needed will simply result in an "efficient information hole."


I also cited a company, publisher Meredith Corp., which recruited "super users" to train on the more advanced features of a new business intelligence system and then share them with colleagues. Users create about 95 percent of BI reports at that company.