We've often discussed the role of failure in innovation - and have started running FailureFests and other devices to get practitioners talking honestly about what they learned from things that didn't work.
We all know how hard this is.
There's a new book out by the guru of failure in engineering, Henry Petroski: To forgive design: understanding failure. He argues that the best way of achieving lasting success is by understanding failure and that a single failure may show 'weaknesses in reasoning, knowledge, and performance that all the successful designs may not even hint at'. For him the best examples are collapsing bridges. Here's a very different, but helpful, example of trying to extract some useful lessons from a well-intentioned project that didn't quite work in a field very distant from bridges. It's a reminder of why it's so important that the new What Works centres are brave enough to set out clearly the ideas that they think have been tested and shown not to work - that may be just as useful as the recommendations on best or proven practice.
Of course it's not enough to say we should celebrate failure. No organisation or system can do that. Instead there is an unavoidable ambiguity in the relationship between innovation and failure. On the one hand if you're not failing often, you're probably not taking enough creative risks. On the other hand, if you fail too much don't expect to keep your job, or your funding.
There is no simple guide to identifying the right quantum of failure (very few organisations have the luxury, or the vast funds, of DARPA which used to aim for a 90% failure rate, and probably achieves something more like a 97% failure rate). Instead, this is something we have to judge and negotiate. But we certainly need more devices like FailureFests to help people talk openly and honestly about what didn't work as well as what did.
And because failure is an unavoidable concomitant of experiment and innovation, we need to get better at failing small and quick. Governments tend to do the opposite. And when we're told that big banks are too big to fail it's surely obvious that the real problem is simply that they're too big.