Why you might need to torture data

TORONTO – Although an enterprise resource planning (ERP) system is great for integrating data across a large organization, this type of software is not well-suited for producing predictive analytics, said a SAS data expert at a recent lecture at the Richard Ivey School of Business in Toronto.

As the amount of data produced by businesses has expanded, the traditional response has been to produce more reports. But according to Mark Moorman, data expert and advisor to Office of the CTO at Cary, N.C.-based SAS Institute Inc., there “is way too much data and not enough information about that data.”

Reports tend to conceal important factors like economics, geography and meteorology, he said. But predictive analytics (along with an assessment of how accurate it is), can add needed value to an organization and help it make better business decisions.

Furthermore, not everyone within an organization wants to be made into an analyst by having to go through a plethora of reports that an ERP system produces, said Moorman. “What they really want is the answer.”

The lecture, the first in a series by SAS and the Richard Ivey School of Business to help Canadian executives deal with today’s business challenges, focused on the advantages of predictive analytics to the business.

But before a business can begin to reap the benefits of predictive analytics, some buy-in may be required. Getting commitment from the executive level could be a challenge especially if the organization has always based decisions on intuition, said Moorman.

But the IT department could be another hurdle considering the tradition of controlling, managing and keeping data intact and “now we are asking them to be a little more free, and be a little more loose, and allow us to torture that data a little.”

IT has historically been in the business of producing ad hoc reports, but when it found itself inundated with requests, IT provided users with tools to run queries on the data, Moorman explained. When that approach, in turn, inundated the user with data, IT resorted to providing alerts on particular data. But the business wants predictions, not reports and alerts.