CAMBRIDGE, MASS. -- The term big data, and some of the general concepts that define it, may be on its way out, but predicting what will replace it and when remains a challenge.
At the MIT CIO Symposium, there was near unanimous agreement that the accepted big data definition, despite its ubiquity, has little meaning for businesses, and focusing on the idea rather than specific business needs may be dangerous.
There are new technologies coming on the market so fast and if we don't really align our organization we can get ourselves painted into a corner pretty quickly.
"The term is useful at some level, but I think if you can't turn it into business value, you're losing," said Darrell Fernandes, CIO of strategic investment products and data at Fidelity Investments. "Big data as a term doesn't connect to business value. The term can hurt us at times if we can't derive the value that we want."
Fernandes said he prefers to talk about "eKnowledge," which refers to information derived from data analysis that connects to business functions.
The debate encompasses more than just nomenclature. Popular use of the term big data came into vogue around 2010. At that time, businesses started to see an explosion in the number of tools available to help them manage and analyze data, including new database technologies like NoSQL and Hadoop, advanced data mining software and, most recently, user-friendly self-service analytics tools. As vendors tried to define big data through their own unique products, it created a data management and analysis ecosystem in which every action related to data requires its own tool.
Fernandes said this can be frustrating. Prior to the growth in big data and its related tools, it was much easier to know what his company needed. Mostly this involved a simple relational database. But now, with so many new tools coming to market, he feels forced to reassess his business's needs on 18-month cycles or shorter.
"There are new technologies coming on the market so fast and if we don't really align our organization, we can get ourselves painted into a corner pretty quickly," he said.
But as the market for data management and analysis tools matures and moves away from general interest in big data, Fernandes expects to see a lot of consolidation. Individual vendors may offer tools that manage data effectively on the back end while also providing easy-to-use data exploration and visualization tools on the front end. Currently, businesses would be hard-pressed to find one vendor capable of delivering on all these needs.
This consolidation could alter the shape of the data analytics workforce. Currently, businesses that want to do deep analytics need at least one data scientist on staff. Doctorate-level data scientist Puneet Batra, founder of organizational learning startup LevelTrigger, formerly served as the lead analytic scientist at Aster Data (which was eventually bought by Teradata), said the need for data scientists like himself is coming to an end. As self-service tools become standard components of vendors' analytic software packages, businesses won't need people with backgrounds in statistics and the scientific method. These things will be baked into the technology.
"My first priority is to put lots of people like myself out business," Batra said.
How long it will take to complete this transition from a fractured technology market to a more unified big data definition is anyone's guess. Tom Davenport, fellow at the MIT Center for Digital Business and professor at Babson College, said he predicted at the Symposium last year that the term big data would have fallen out of use by now -- a prediction which obviously has not come to pass.
While big data as a general concept may provide little business value, as a term it has proven durable.