News Stay informed about the latest enterprise technology news and product updates.

No single path to building big data analytics platform

Making IT an integral player in your company's big data analytics platform calls for out-of-the-box thinking about IT's role. Here's some advice on how to get there.

As companies become more reliant on technology, experts have encouraged CIOs to recast their relationship with the business -- innovate, embrace Agile practices and experiment with bimodal IT. But to get there, CIOs will face a steeper hill than sheer departmental transformation. They'll have to convince the business they're entitled to change. If the recent Big Data Innovation Summit in Boston is any kind of litmus test, there's no easy prescription for adapting IT to digital business.

"In large, enterprise IT organizations, the mandate is really about, 'don't fail,'" said Moin Haque, senior director of the analytics technology group at NBCUniversal Media LLC, during a panel discussion at the summit. "They're measured by not failing; they're measured by [service-level agreements]. That really does not give you the right space in which to take risks and fail fast." Haque's analytics team, on the other hand, operates under a polar opposite mandate: Take risks, move fast and fail fast, he said.

NBC's big data analytics platform is a good example of IT's bifurcated role at the network: The IT team manages the in-house Hadoop cluster; Haque's team, which is only a year old and exists outside of IT, experiments, iterates and innovates with data in search of business value. Because the two teams operate at different speeds and are tasked with different goals, Haque considers the division of labor a benefit. "Taking that responsibility out of our group is one of the advantages," he said.

Outsource elements of big data analytics platform

Haque advised other businesses follow suit when starting down the big data analytics platform road. "That separation, if it can be done in-house or by a partner, that is something we would recommend as a starting point," he said. In fact, when an attendee asked the panel whether to outsource or insource a big data project, he recommended the attendee break it up, separating the big data analytics platform from other aspects of the project.

The platform, he said, could be turned over to IT to manage, or, because the organization doesn't have big data skills in house, "maybe you outsource that," he said. Big data analytics, on the other hand, should be outsourced only rarely.

"Whether it's developing models or algorithms, you really want to keep that in-house, because that, to some extent, is going to be proprietary and is going to be competitive," he said.

Cultivate in-house talent

Panelist Chris Simokat, vice president and lead data scientist for the big data and analytics engineering team at Citigroup Inc., described a model that could serve CIOs well. Rather than outsource platform duties because of a lack of in-house skills, CIOs should grow internal talent. "One of the origins of our own platform -- the core engineering component of it -- is actually a spin-off from our PeopleSoft engineering group," he said.

Simokat echoed Haque by suggesting businesses break up big data projects into smaller, more manageable pieces. Then, look inside the IT organization for employees who can become in-house experts for each piece of the project, even for the work that gets outsourced.

"Otherwise, you're at the [mercy] of the contractors, and you're going to have difficulty validating what they did," he said.

The innovation paradox

CIOs and businesses alike say they're hiring creative and innovative people, but talk is cheap. "Most innovative people are a pain in the neck because they like to change things," Atti Riazi, assistant secretary-general and chief information technology officer at the United Nations, said during the summit. "No government wants them, businesses don't want them."

Innovation is a double-edged sword, Riazi said. It can be used to solve the big, intractable problems, but it also can bring unintended and often unpredictable consequences. Take the Internet. "It's something we cherish, we love," she said. But the Internet is also home to the "Dark Web," an encrypted network that cannot be indexed and is rife with things like human trafficking.

That in no way means CIOs should shy away from innovation, because "without innovation, we're doomed," Riazi said. Instead, she encourages attendees to start talking about -- and thinking about -- technology in a way they may not be used to. "We think technology is about stuff -- hardware, software, tools, cloud," she said. "But it's much bigger than that -- it's about experience, power, the impact on human life and the environment. That's where we should be having conversations."

Riazi said that trying to manage innovation and technology is a mistake that will "choke its vitality." The only thing left to do? "Innovate in a responsible way," she said.

Let us know what you think of the story; email Nicole Laskowski, senior news writer, or find her on Twitter @TT_Nicole.

Next Steps

Read more from the Big Data Innovation Summit here.

Dig Deeper on Enterprise business intelligence software and big data