By Robert L. Scheier
A lot of outsourcers are making noise about handling big data projects for their clients. In some ways, big data – the analysis of very large datasets for business insights – seems a good fit for service providers. They have the skills and, in some cases, the infrastructure needed to handle the exceptionally large quantities of unstructured data (including text and video) their clients need to analyze.
But a look under the covers, courtesy of a session on “Big Data and Analytics” at the 2012 MIT Sloan CIO Symposium, unveiled some unexpected twists and turns. Many of them revolve, not surprisingly, around human behavior and corporate culture as much as technology. They include:
• While big data involves heavy technology (specialized databases, grid computing, visualization tools) it often isn’t owned by IT, but by business areas or even R&D, said Professor Tom Davenport of Babson College. My suggestion: Service providers need to remember this so they’re fishing for business in the right ponds, and bring their vertical market specialists, and not just their geeks, into the sales cycle.
• Clients are spending – or wasting? – a lot of their time on data management (finding the data, categorizing it and cleansing it of errors) instead of actually analyzing it, said Davenport. The possible opportunity: That service providers may be able to handle this lower-level work, leaving customers to ponder more business-critical analyses.
• One very human problem is companies not accepting “big data” results that don’t fit their preconceptions, said Shvetank Shah, executive director of the Corporate Executive Board, sticking with their original hypothesis “even when the data disproves it.” Davenport recounted the classic finding that beer and diapers are the two items that are purchased together most often in convenience stores. But he went on to say the retailer never changed its store layouts as a result because they didn’t trust the results. A service provider who’s trusted by the business leaders (and has wider expertise in the client’s vertical) could help fight for at least considering such unexpected, and possibly valuable, insights.
• The more data, the better, at least when you’re trying to uncover, say, side effects from various drugs or how strong the holiday shopping season will be. Many customers are wondering whether, and how, to share data with partners and competitors to get more accurate results without jeopardizing security. While data sharing could drive huge reductions in health care costs, for example, providers are held back by privacy issues, said James Noga, vice president and CIO of Partners HealthCare. The right service provider, or group of providers, could possibly serve as the “neutral arbiter” to set standards for data sharing and even handle the grunt work of sharing and managing it.
As the older saying goes, “information is power” and the answers uncovered by big data can be powerful. Service providers, and customers, who outsource big data will have to keep an eye on far more than the “volume” questions of which database computing platform to use. Who owns the data, how it is shared, and how to make sense of what you find will be far more difficult issues to solve.