Financial Executive Insights

Playing With Big Data

Laura Schreier | Special to

Newton, Mass. — Experts on Big Data at the MIT Sloan CFO Summit on Thursday shared their thoughts on how to implement the technology, how to use big data, and what kind of privacy concerns it raises.

Dave and Les Jacobs | Blend Images | Getty Images

But first, they had to explain what it is.

Although the concept is surrounded by a lot of hype, Big Data is also not well-understood, even by many sophisticated companies. Panelist Ron Gill, CFO of NetSuite, started the discussion by explaining that, with the rise of social media, people now have a ton of forums to express their opinions on products, companies, even political candidates. Before, a consumer might have had strong opinions on a restaurant — but now they can Yelp about it.

Businesses can pull together that type of data and use them to make smarter decisions. Big Data is usually talked about in relation to social media, but it can also be any large set of scattered data points that can be drawn together to make some kind of conclusion.

But this data is "unstructured" – randomly dispersed and hard to collect – and there is a lot of it, said Jeanne Johnson, principal of business intelligence for KPMG.

"The biggest characteristic about Big Data" she said, "Is that it's big."

Given that, it can be a challenge for companies to know where to start. Justin Borgman, CEO and co-founder of Hadapt, encouraged a slow approach. Make a small investment, link a network together and "start playing." Hire somebody who knows how to use these technologies and talk about the possibilities. Borgman compared it to psychology's "Maslow's Hierarchy of Needs," where companies start at the very basic, foundational questions, and only later culminate onto actual Big Data projects.

But Andy Palmer, director of CloudSwitch, added a warning to companies looking for Big Data vendors.

"Caveate emptor," he said. "The number of new technologies funded by VCs – it is not just extreme, it is irrational."

Some of those companies are great. But many are purely raising money from VCs, not doing a lot of development, and then turning around and selling poor product and poor ideas to unwitting clients.

Fortunately, panelists said, this industry is developing quickly. The engineering is progressing fast, and it is becoming easier for companies to integrate Big Data into their businesses.

As with any project, companies have to spend time in advance thinking about what they want to accomplish, Gill said. If they figure out what questions they want to ask, the infrastructure and engineering are developing fast enough to help companies answer those questions.

Making Big Data pay off doesn't require massive effort, Gill said. Once a program is in place, it can be easily integrated into business as usual.

Gill described one example concerning a retailer that sells sports jerseys. This retailer always tracked store data, he said, looking at which locations sold out of which types of jerseys, and then shipping extra to those locations.

Now, with Big Data, it tracks every major league sports players' name on Twitter, watching for wherever that name has a positive spike.

Tim Tebow would be popular in New York, for instance, but the business found that the player is popular in the Deep South because of his well-known Christian beliefs. The retailer could immediately ship more Tebow merchandise south to meet the demand as it rose, instead of waiting to track sales data months later.

Doing that, Gill said, required no major operational change – just a newer, better report to look at.

But Big Data comes with risks, too.

"This stuff can start to get pretty creepy, pretty fast," Gill said. For instance, some vendors currently offer to help companies take every customer email address in their records, identify the Facebook or Twitter handles associated with it, and compile a report that tracks every mention of a relevant product or opinion, delving deep into that customer's activity.

Johnson noted, however, that the risk is likely to be reputational, rather than legal. And context is important – for example, Amazon collects a huge amount of data on its customers, an amount that could easily be seen as too invasive, except Amazon customers don't object because they are knowingly handing the company this information.

Big Data, however, continues to move fast. Panelists noted that the technology is advancing to capture data from previously unheard-of sources.

Borgman mentioned a speech-to-text program that captures spoken words and analyzes them, and Palmer had heard of program that is able to analyze a video stream on the web and identify whatever brand names show up.

"If it can be produced, it will be captured at some point," Johnson said.