A nice article yesterday from Aaron Erickson about the relationship — I would go as far as to say the symbiotic relationship — of agile software development and failure. He provides an overview of methods to identify a failing project and metrics that can help discern failure as quickly as possible. To quote:

Failing fast doesn’t just save us money. The cold reality is that people don’t like working on failing projects. When a project is failing, people generally can smell it in the air. Employee engagement goes down, turnover goes up, and things just generally get worse. The best thing to do — for your people, for your company — is to learn to fail, and create a culture that learns from failure.

I heartily agree, though I would emphasize that failure doesn’t have to relate to the entire project, and ideally it doesn’t. You could fail fast on an architectural approach, a feature, a graphic design, or an idea. As Alan Cooper recently wrote:

Generally, ideas are cheap and plentiful, therefore so are good ideas. Unfortunately, bad ideas are the most plentiful of all. While success depends on developing good ideas, consistently successful companies and individuals are skilled at both identifying and discarding bad ideas to devote resources to the good ones…. You have to be tough enough to nurture your idea to the point where its fatal flaws become visible, then ruthlessly put it out of your misery. The author William Faulkner said, “Kill your darlings.”

Good night, darling.