Good innovations need good failures

In mid February the Churchill Club ran a program called Innovation DNA, which was about making innovation a sustainable program in Australian SMEs. 

One of the panelists, Wes Sonnenreich from Deloitte, mentioned that a good failure was one that was quick and cheap.

Which got me thinking about failure again.  My assertion has been for a long time that Australians in business are generally terrified of failure, as it creates a stain on your reputation. Which is fascinating, as you can’t have reward without risk, and risk means there will be failure.

An interesting contrast then, to the business approach, is how Australian science treats failure.  Failure is absolutely a necessary part of any experiment, and just as important as success.  For example a test for a link between cancer and smoking needs to also prove that there is no link between sucking on a tube and cancer (a failure).

The scientists understand the importance of failure, and also understand how to give it good PR.  Most grant applications are to test whether there is a link between two areas, instead of trying to prove there is. 

The science approach to grants means that you always have a successful outcome (either you proved there was a link, or you proved there wasn’t a link).

So back to Wes’s comment about good failure being “quick and cheap”.  It seems to me that at a qualitative level, a good failure is one that you can learn from (which is why we do exit interviews when employees leave).  But it still leaves a problem on how do you quantitatively measure the value of failure.

Well, Wes working for Deloitte gave me the clue.  It made me think how would an accountant treat the problem?  Although it was dusty and unused, I pulled the accountant hat out of the cupboard had a go at it.  My thoughts went like this

The value of an activity is normally measured by an accountant with a profit and loss statement.

2. Failure generates costs.
4. An innovation program will have a budget that’s used to generate lots of failures.

Cutting to the chase, here’s the idea. Why not create a profit and loss statement to measure whether its a good failure or a bad failure.

The first thing you would do is apportion the budget of an innovation program across a number of projects.  This would give you a notional revenue figure for each project.

You then apportion the costs of each project to its project P&L.

If the project fails quickly and with low costs, you will get a notional profit, which could be interpreted as “good failure”.  If you take forever to fail and commit lots of resources and get a notional loss, it means it’s a “bad failure”.

Obviously this idea needs to be tweaked for the situation, but I feel it would provide a great platform for starting to measure the value of failure, and give it better PR in commercial environments.