Bullish on big data? Better make a plan.
A well-reasoned strategy is essential for big-data success.
Big data efforts are likely to fail without a well-reasoned strategy. (Stock image)
Are you bullish on big data?
Then your agency better get a big data strategy in place – and if that strategy does not outline at least one specific problem to solve, a timeline for success and measurable metrics to gauge progress, it’s probably a waste of time and taxpayer dollars.
While federal agencies fall at different points on the big data learning curve, the intelligence community (IC) sits comfortably at the top of the class, in part because they’ve solidified what they want from big data and taken steps to make it happen. The CIA, for example, recently signed a $600 million deal with Amazon Web Services for the construction of a private cloud infrastructure that will almost certainly provide a direct boost to the CIA’s big-data efforts.
The collection of 17 agencies in the IC have made it clear through strategic actions and occasional public statements that they are embracing big data – massive stockpiles of information produced from sensors, social media, personnel records, smart machines and other data producers – to analyze it to find as many needles as they can in exponentially growing haystacks of data.
"The general premise that the IC and Department of Defense are ahead of the curve is accurate for understandable reasons," said Chris Wilson, VP of Federal Government Affairs for TechAmerica and staff director of its recent Big Data Commission.
"They’re at the top, but there are other agencies I think that are seriously looking at big data and putting together a game plan and using it already to varying degrees of success," Wilson said.
The departments of Energy and the Treasury, as well as NASA, the United States Postal Service, and several sub-agencies within the Department of Commerce have all produced innovative – and sometimes very cool – big data success stories.
"Most of the others are still in the very preliminary stages of putting things together," Wilson said.
'Lack of adequate resources' leads the way among reasons agencies don't do well at data collection and analysis. (Government Business Council graphic)
Why a strategy?
A recent survey by the Government Business Council suggests big data strategies are "the defining element to unlocking the potential" of growing mountains of information. The survey, underwritten by Booz Allen Hamilton, polled 313 executives in 27 federal agencies and reported that only 37 percent of managers reported their agency is taking "appropriate steps" to leverage big data to enhance agency operations.
Furthermore, only 31 percent of those surveyed felt their agency was "fully leveraging all of the data it collects," and equally troubling that only about one-third of those polled felt their agency was properly using technology to expand their data capabilities or investing in their current IT workforce to learn new skills. Perhaps the most striking figure in the survey was that for all the talk in the past year in the federal IT community about the importance of data scientists – the analysts who turn data comparisons into insights – only four percent of those polled said their agency was hiring these scientists.
Despite causes for concern, 69 percent of the executives polled said they felt big data had the potential to fundamentally transform federal operations. Yet getting from point A to point B isn’t going to happen without hard plans in place, according to Mark Herman, executive vice president at Booz Allen Hamilton.
Herman is outspoken about the importance of big data, though he hates the term itself, preferring instead to call it multi-disciplinary data, which he said encompasses rich sets of data that come from more than one stovepipe database.
A good big data strategy, he said, should begin with an outlined path of "where you want to go," and should be followed by a set of requirements based on an agency’s mission, kind of like a how-to kit for pulling value out of a big data investment. Budgeted dollars, staff commitments, personnel and potential technology procurements and outcomes would ideally fall under this strategy.
"This is a transformational process, it’s deciding who you are and how you operate and understanding that when you go down this path, you’re going to be different than you are now," Herman said.
What a strategy doesn’t have to be is expensive, he added.
That doesn’t mean big data platforms don’t cost money – they most certainly do – but a proper big data strategy is going to include how older legacy IT systems are replaced or utilized less, while more investment goes toward newer technology.
That part may be difficult, Herman said, because there are "a lot of politics around how the government can and does spend money," but continued pressure from top agency IT executives can spur that kind of change.
Disagreements may form between cost-conscious CIOs tasked with meeting tight IT budget requirements and operating executives focused on successful mission implementation, and how those are handled could very well determine the speed at which big data technologies are adopted by federal entities.
Herman also touted pilot programs as intelligent decisions.
"Don’t try to fix the past, that’s expensive – fix the future," he said. "If you are an IT guy in a government agency and you still have a developmental budget, instead of spending money to buy more of the game, use it for new things around concepts like data lakes, cloud structures and building the next generation of innovation and try to drive value from it."
NEXT STORY: Join the Conversation about Federal IT Reform