Failure is absolutely critical.

I think it is not possible to innovate without failure. If innovation is the process of taking significant steps forward, steps past where you can't see for sure what the conclusion will be, you can't possibly not fail the bulk of the time.

There are many sayings that our society has for this, such as “If you're not falling down at least occasionally, you are not skiing fast enough.” Now the question is not whether you need to fail. You do need to fail a decent amount of the time, or you can't possibly be an innovator.

The question is: Does our culture as a society, and does your culture at your particular company reward you for taking good risks? Stupid risks, you should be punished for. It is whether you took a good risk, not whether you succeeded or failed, that matters. Because you can't know the outcome ahead of time. If you are going to be punished for failing, then you're just not going to try new methods and test hypotheticals.

http://curiosity.discovery.com/question/failure-okay