Scott Cook: Innovation Fueled by Experimentation

There's another piece I'd add to Fostering Counterculture. When you get a grand challenge like that, then people come up with wildly different ideas on how to achieve it. Ideas that most people would say are borderline impossible. So then what we do is let people run experiments. Because in a company where you have to get everything approved, then only the people who are most persuasive or have the most tenure or the highest ranking title will get their ideas approved. In a company where you run on experiments and decisions made by experiment and you make it really easy for people to run experiments on their idea, then the idea that will get approved is the idea that has an experiment proving it worked. Now, I didn't invent this. I first observed this at Toyota, where I went over and visited Toyota plants with professors who'd studied them. And they explained to me that you have to understand the company runs on a nested series of experiments at all levels. And they described one experiment that was on the line and they were running, trying to change the production process with an experiment in two different ideas. One idea was the plant manager's idea, the boss. The other idea was from a supervisor who manages a team of like eight people. Now, in a US auto plant or German auto plant, it would have been the boss's idea and they wouldn't run the experiment. But in Toyota, they ran both experiments to see which would win. So I've kind of come to a point of view that leaders have a few key roles in innovation. One is to declare the grand challenge, what is that world changing problem you're going to try to solve? Then two, set up a system for people to run experiments so that new employees can run experiments and try their ideas out and make it fast and cheap. So the company runs a lot of them. Google in search runs something like three or five thousand experiments a year in search. Then savor the surprises, because in those experiments you get surprises. A bunch of them don't work like you think. Everyone that doesn't work is an opportunity for learning. There's something you didn't know when you form that experiment and set your hypothesis. And the experiment is telling you you didn't know. Now figure it out. And the fourth thing for the leader is live by those same rules yourself, so that your ideas, unless it's a company emergency or something. But in normal business, your ideas have to be tested as well. Just like that Toyota plant where the boss's idea was subject to the same test as the idea from the line supervisor.