Check out Fast Company’s thought provoking Q&A with Roger Martin who explains why “you cannot prove a new idea in advance by inductive or deductive reasoning.” Martin cites A.G. Lafley at P&G as a rare example of a CEO who was able to look at the data provided by analysts, and then push it aside:

Martin: When he first took over, A.G. Lafley at P&G was brilliant enough to realize they were missing a lot about the holistic consumer experience by sticking to things that were rigorously quantified. For example, when the company moved into beauty products, they were looking at face cream. And the scientists decided it must be about pore coverage. So they analyzed the hell out of pores and said ‘We can cover pores better than anybody.’ So when women in their research started talking about wanting to feel beautiful and desirable, they’d say, ‘Don’t talk about that. We don’t know how to quantify that!’ And they couldn’t understand why stupid women would go off to department stores and pay ten times more when they could cover pores just as well. Ten years ago, P&G couldn’t prove they could sell women billions of dollars of Oil of Olay face cream at $30-$60. They could imagine it, but not prove it. Lafley took it as a management challenge to see across the divide.

Fast Company: If you don’t have A.G. Lafley or Steve Jobs at the helm, how can you sell your organization on the idea of an intuitive leap instead of a scientific leap?

Martin: You don’t have to convert the whole organization to design thinking. Propose a little experiment–say, three months in length–where you test out a bite-sized chunk of a problem using this method. If you have a little success, be sure to then attach metrics to it. In that way, you turn the future into the past in a way they understand.

Click here to read the full interview.