Better Decisions with Smarter Data

While the concept of information overload isn’t a new one — Alvin Toffler introduced it back in 1970, in his book Future Shock — it seems more relevant now than ever. Particularly for the growing number of organizations with a mandate to make more strategic and operational decisions based on data — or facts — in environments saturated with data.

There is so much data in market and non-market environments that it has become a cliché to note data generation and consumption in funny terms like exabytes (sounds like an orthodontist’s call to action) or zettabytes (a word that, for some people, may call to mind toothsome college fraternity zombies). The irony, however, is that there can be both too much data yet too little good data available when the time comes to make decisions.

In our recent data and analytics survey of about 2,500 professionals — part of our second annual research collaboration between MIT Sloan Management Review and SAS Institute — 60% of respondents agreed that senior managers are pressuring the organization to become more data-driven and analytical. At the same time, only 42% of respondents said they “frequently” or “always” have all the data they need to make key business decisions.

To cut through the noise to get the data that is most useful and timely requires smarter data, says Ali Fouladkar, a researcher at the Center for Studies and Research in Management (CERAG) and Ph.D candidate at the Doctoral School in Administrative Sciences within Université de Grenoble Alpes. Fouladkar defines smart data as data from which signals and patterns have been extracted by intelligent algorithms. Imagine the difference between a long list of numbers referring to weekly sales, versus a graph that tracks sales peaks and troughs during the same time frame, and you have the basic idea of what separates ordinary data from smart data.

In his research on data-based decision making, Fouladkar defines three key attributes that distinguish smart data from other forms of data. To be smart, data must be:

• Accurate – data must be what it says it is with enough precision to drive value. Data quality matters.
• Actionable – data must drive an immediate scalable action in a way that maximizes a business objective like media reach across platforms. Scalable action matters.
• Agile – data must be available in real-time and ready to adapt to the changing business environment. Flexibility matters.

The problem in getting to smart data is not just technological, it’s also a training issue, says Fouladkar, in a presentation (and paper) given at EdCon 2013 symposium, a joint effort between IBM, Syracuse University School of Information Studies and the University of Ottawa Telfer School of Management.

Fouladkar, as a Big Data lecturer, is developing a new degree program at the Graduate Business School, IAE Grenoble within Université de Grenoble that integrates strategic management, marketing, finance and management information systems [MIS] programs. He is also proposing the development of executive education courses aimed at helping senior decision makers better understand and utilize data.

IAE Grenoble isn’t alone in its march toward big data training and education for managers and technologists. Last summer, InformationWeek ran its list of the Top 20 Big Data Analytics Masters Degree Programs (MIT Sloan School of Management is on the list).1 And there have been a number of alliances between corporations and universities seeking to prepare students — and managers — to better utilize data in decision-making. IBM, for example, participates in more than 1,000 partnerships with universities across the globe focused on big data and analytics.

But the role of universities and corporations must be substantially expanded to enable executives to make better decisions, and train the next generation of decision makers, says Fouladkar. The current challenge: Most of the degree programs offered currently are based on “pre-Big Data logic” and do not adequately embrace big data analytics, business intelligence and business modeling. “Education and training lack coherent learning objectives,” Fouladkar observes. “Available training was built around technical and engineering problems and it does not offer proper theoretical basis to support decision makers in proper exploration of the Big Data field.”

One solution: more executive programs that teach senior leaders about the potential of smart big data — and its implications for corporate strategy and organizational design. Fouladkar suggests the establishment of university-business laboratories for research and teaching purposes. These joint experimental laboratories would enable universities to conduct more applied research and make it easier for students and managers to prepare for the challenges they will face in their business careers.

Because the thing is, data can’t be the only smart thing in the room.