Making Innovation Happen: 2016 IRI Medal Address

Abstract

discussion is about innovation--about the conditions under which innovation and invention happen. I have a few thoughts on this, from my time at Google, DARPA, IBM, and MCI. True breakthrough innovation, what we call moon shots, requires a specific environment, one that allows failure and celebrates success. Creating a productive platform for innovation--a breakthrough that others can build on--like the Internet or a mobile operating system, also demands attention to some specific elements of technical design, to allow flexibility while ensuring interoperability and to build in resilience and adaptability in the face of unexpected technological developments. Creating an Environment for Innovation To facilitate innovation, an organization has to create the right environment. That environment has some specific attributes: an atmosphere that allows for failure, perhaps even celebrates it, one that encourages interaction and gives people the freedom to explore. One of the key elements of an innovative environment is the freedom to fail. Most scientific experiments actually don't work. We're a little lax about publishing experiments that didn't work, but we could all learn from them, especially if we can explain why the experiment failed. Was it just a poorly conceived experiment? Was there experimental error? Or did the results invalidate a theory? Failure may also be a matter of perspective. Maybe the experiment was mostly successful, but there was this one thing over there that didn't fit the theory. One scientist will say, Measurement error, and ignore it. But another one will look at those results and say, funny. Why is that? That's the scientist who gets the Nobel Prize or produces a pioneering innovation. At Google we have that freedom. We don't have to succeed every time we try something, but we are all asked to shoot really high. The idea is that if you shoot high when you try something, even if you miss that goal, you may still get a pretty significant thing done. freedom to fail is very important. The structure of the organization is also important. Real innovation requires interaction, both within teams and across projects. People need to share ideas. Casual, unplanned interactions are enormously productive and important in this context. At Google, we've found that small teams really help. With a reasonably sized team-- three, five, seven people--a great deal of interaction can happen quickly and you don't spend a huge amount of time bringing everybody up to speed. If you have a larger group, it's much harder to do that. Another lesson we've learned as Google has grown is the value of co-location. The company grew globally and ended up with some teams whose members were many time zones away from each other. That proved to be very inefficient. So we went through a cycle of what we called defragmentation to co-locate people who were working on common problems. That helped with our efficiency, again, because it facilitated interaction. It's also important not to lock people up too tightly in their teams. Pixar, to give an example, wrestled with this. To make a movie, production companies actually create a new company for each movie; people come into that company just to work on the movie. At an organization like Pixar or Skywalker Sound, which works on other people's movies, the people working on a given project are typically compartmentalized; they're not allowed to talk to anybody else. Everything is secret. Steve Jobs looked at this arrangement and said, roughly, This is stupid. We have really, really smart people in this company and I don't want all of them compartmented and separated from each other. If I have a problem with movie X, I want to have access to every bit of talent in my company. He opened it all up so that anybody at Pixar could work on or be drawn into any project. That meant every movie, every project, could take advantage of the full range of expertise in the organization. …