“Why Your Analytics are Failing You.” That’s the provocative title of business guru Michael Schrage’s latest post on the Harvard Business Review blog.
Schrage, prolific author and research fellow at MIT Sloan School’s Center for Digital Business, makes a critical and often overlooked point about what it takes to extract full value from analytics initiatives.
It’s not the data, it’s the behavior
According to Schrage, in the real world, it’s not just a matter of having the right data, the right analysts, and the right tools, nor is it about the chicken-and-egg question of whether to capture as much data as you can up front or collect only the data needed to answer specific business questions. Instead, Schrage believes the determining factor is how the analytics are used.
“Companies with mediocre to moderate outcomes use big data and analytics for decision support; successful ROA—Return on Analytics—firms use them to effect and support behavior change. Better data-driven analyses aren’t simply “plugged-in” to existing processes and reviews, they’re used to invent and encourage different kinds of conversations and interactions.”
If you’re just using analytics to improve what you’re already doing, you’re limiting yourself to incremental improvements. Valuable, sure, but seldom transformative.
Are you willing to change based on what you learn?
The big payoff comes for companies that recognize “that new analytics often requires new behaviors.” Existing business processes and incentives usually have to change to reflect the new information.
Schrage argues that the quality of big data and analytics may matter less than the purposes to which they’re put. The questions and answers derived from analytics remain important, but the key to being a high-ROA enterprise, Schrage claims, is making sure the results align with individual and institutional behaviors—and vice versa. Companies that aren’t willing to change their behavior in light of what they learn from analytics may find they don’t get the benefits they expect. In many cases, he says, it’s not that analytics fails the enterprise, but that enterprises fail their analytics.