Think about how much it costs to learn more. Sometimes you want to build confidence by knocking off the easy things.

If you are a large company and you want to do something unique, you almost by definition have to tap onto the core business in some way. Otherwise you are going into a naked fight against startups, and that's just a tough place to be.

It's one of the underappreciated skills required by an innovator - they have to be able to convince lots of people to do things that might not be fully rational (invest in the company, join something that is likely to fail, try a product they've never seen before), and if you can't tell a good story it is just very hard to make that happen.

You have to chip away at it and shape it to let the winning idea emerge.

When we are looking at ideas, we really are thinking of three things. Are you doing something that seems consistent with the patterns of success? Are there reasons to believe you have or can access the right people to make it happen? Can we play a unique role in enabling success? We all have our rules of thumb, I suppose.

You can always figure out how to deliver things in somewhat controlled situations, but when you start to get into the reality of the market you start to figure out what isn't going to work.

The need to be thoughtful about experiment design is particularly acute within large companies, since some of the behaviors, such as having small teams and tapping into low-cost resources to maximize flexibility, won't come naturally to many people inside huge companies.