When I hear it said that we should celebrate our failures or that we should “fail fast”, I fear that some people may mistake the intent behind it. The aim is not to fail, but instead to learn from failures. What we should really say is “learn fast” and “celebrate our discoveries”.
As a fan of many forms of motorsport, I was interested to read about a recent experiment with adding “mudguards” to a Formula One car. (The aim being to reduce the spray thrown up by cars travelling at high speed in wet conditions.) One article caught my eye because its headline declared “First F1 ‘mudguard’ test ‘a failure’ – report“. My immediate thought was that the test probably wasn’t a failure – far more likely was that the test didn’t show it to be 100% successful which, given it was the first test of a new concept, would have been unlikely. The report said that “there was still too much spray” but that “The test provided valuable CFD correlation data as well as good driver feedback.” (CFD in this instance means Computational Fluid Dynamics rather than a Cumulative Flow Diagram.)
My takeaway from the article was that data from the test will help them improve the design and their approach for the next iteration of the device. Is that a failure? It didn’t solve the problem (how often does Version 1 of anything do that?) but I would only call the test a failure if they came away from it with no new information.
Rather than call it a failure, why not say the test was successful in generating further insights? Probably because that’s too long for a headline, and “bad news sells”. But when we’re talking about running experiments, especially when we know status reports are summarised as they go up the food chain, why not say the test was successful or that we now know more than we did before the test? Mislabelling it as a failure could lead to unfavourable repercussions, including senior managers wanting to stop experiments because “we don’t have time for failure”.