I wrote last week about the challenges involved with setting up centers of innovation within large enterprises. Yesterday I attended a breakfast at the Stanford Faculty Club and our guest speaker was the world-renown Steve Blank. Steve is undoubtably the smartest guy I know on this topic.

Big companies today sometimes look at small startups to find best practices for innovation. But the problem is that startups are not simply small versions of big companies. They are really very different animals.

As Steve is found of saying, "Large enterprises exist to execute business models; startups exist to find a business model". Those are two very different mandates, and the structure and tools required for execution are quite different from the structure and tools required for innovation. 

The most famous example of successfully creating a center of innovation within a large enterprise is the Skunkworks which was started within Lockheed during the cold war. By being given a high degree of autonomy, unhampered by bureaucracy, this small group inside a huge corporation turned out some of the most advanced aircraft designs in history. Innovation within a $45 billion corporation ain't easy, but the Skunkworks did it. 

At breakfast yesterday, Steve offered a few tips for large companies wanting to set up their own internal innovation centers:

• Staff the innovation center with cross-disciplinary teams. Engineers will come up with things that no customer wants, so put salespeople and product people on innovation teams if you want the work to be informed by real world needs.
• Small wins are important. Create a culture of experiments and prototypes. 
• Speed and urgency are fundamental to innovation. Tell the team you are going to shut down their center of innovation if they don't come up with interesting things in the first six months. You'll be amazed that the innovation that happens when you light a fire under someone's ass.
• Free them from legacy constraints. Don't let the IT department dictate tools and platforms. Remember, innovation and execution are completely different things – innovation only happens in the absence of legacy constraints.
• Agile engineering processes are essential. Among other things, this means taking a "fail fast" approach. If a project doesn't have promise, kill it and move on.

As always, I enjoyed Steve Blank's thoughts at breakfast yesterday. Creating centers of innovation within large companies isn't easy, but and essential part of making sure companies today remain relevant in a fast-changing marketplace.