- Unless you can show that your project will make the company more money in the first couple years than putting it into the bank it's a no go. We're told that it's cheaper to wait for another small company to succeed and then just go and buy that.
- We have IP, great labs, technicians and vendor relations for the technologies we're in already but it's hard to expand into totally new areas because we wouldn't be utilizing the existing resources. The usual answer is outsourcing but it's very hard to find good help when you don't quite know what you're doing yet.
- We had an official innovation funnel for a few years where employee ideas could gain some traction but that approach was abandoned in favor of the few-big-bets approach. I can't honestly say why the funnel failed but after a short while we called it the "funnel of death" since the financial types would never accept the business plans. At some point we tried to keep our side projects out of the funnel hoping we'd get far enough on our own.
We make mostly hardware so we're a lot less nimble than pure SW shops but some of the same trends surely apply.
This is probably not unlike what actual startups run into, even after the lessons from the dot-com bust: there's a lot of enthusiasm around the technical ideas but it's hard to convince investors.
And in a similar vein, organizations used to large-scale upfront requirements probably aren't even well-positioned to adapt to a new idea that comes out of a failed project. (Think Twitter being born while working on Odeo).
I'd be really interested to hear about examples of large companies that take this approach.