Netflix had a singular focus when designing its innovation competition—deliver a more accurate prediction of a customer’s movie ratings in order to improve its recommendation engine. Now, the company says it realized predictability is just one of the many components of an effective recommendation system. And Netflix has decided that, due to the complexity of integrating the algorithms into its existing systems, and the company’s changing business landscape, it won’t implement the algorithms that won the Netflix Prize. (For more, see “In Awarding Prize for Analytics, Netflix Failed to Predict It Wouldn’t Be Used.”)
The problem for many companies seeking innovations driven by big data is that they’re not asking the right questions, says Michael Schrage, a research fellow at MIT Sloan School’s Center for Digital Business.
Schrage says business leaders need to ask these questions when embarking on an innovation project:
- Who are your best customers? How do you know?
- Who are your typical/average customers? How do you know?
- What are the two most significant differences between your “best” customers and your “average” customers? How do you know?
- What are the two most important things you can do to create new value for your best customers?
- What are the two most important things you can do to create new value for your average customers?
- Why is or isn’t there overlap between innovation for your best and most typical customers?
“Designing analytics and metrics to answer those questions is what serious Big-Data-driven innovators should be doing,” he says. “[But] only a fraction of organizations have the self-discipline, rigor and honesty to actually act on these questions.”