Fail Fast Fail Cheap – Is PDCA always better than controlled trial and error?
I heard an interesting debate on the radio the other day. Two academics were arguing about the impact that the information age has had on us and the way we do business. One academic said it has been bad for us and (you guessed it) the other argued the opposite. The debate had clear implications on the way we may approach quality management. I’ll start the article with some background to the argument.
The information age is bad for us – the case for the prosecution
So, back to this radio debate, the academic who argued that the information age had been, on the whole, bad for us, had a good reason for saying it. He stated that in the past, when we had to spend a week painstakingly researching something, we valued what that research eventually yielded. It would have constituted a significant investment for us as we’d have to put in our library hours and, because of that, we’d want to extract as much value from the result as we could. He argued that the fact we can find out in 5 minutes what it used to take us a week to uncover, causes us to undervalue the outcome and we often don’t bother to learn from it, or even process it properly. This is certainly something I can identify with and, to a degree, recognize in myself. But, as always, there’s a flip side, and here it is ….
The information age is bad for us – the case for the defence
Old business rules no longer apply. Things move more quickly, things are more transient, life-cycles of products and of trends are shorter. “Right first time” is dead. We haven’t got time to get it right first time. By the time we’ve got it right, the opportunity has often passed.
This change in dynamics brings two implications;
- We need to be quick, agile to seize small windows of opportunity and, (as a result of moving fast);
- The incidence of defects and failures will increase
Now, as quality professionals, that gives us a problem. Failures, we are taught, are bad and should be avoided at all cost.
But hang on, that’s not quite true is it? Expensive failures are bad. Damaging failures are bad … but if we can find a way to fail fast …. and fail CHEAP … then …
So “Fail Fast Fail Cheap” is the new PDCA. Right?
Well, not quite. In some industries cheap failures are difficult to manage. Food, drugs, medical devices, aerospace? The idea that these industries could evolve into a fast moving trial and error model is inconceivable. Their failures are potentially dangerous and very damaging. The chances of a quick and graceful exit after a failure in those industries are small. So PDCA certainly is not dead. But fast moving consumer markets with low public safety risks … that’s different, and we need to accept that different, and perhaps mildly sacrilegious, rules may now apply. If controlled trial and error is the quickest, easiest and cheapest option, why would we do any other?
Anyway, for those of you looking for some new rules, I hope this is a new one worth thinking about. Here’s an article on the subject I found in Business Week on the same subject. However latterly it is important to understand the context of a process of “controlled trial and error” in relation to the management of risks and opportunities. Absolute certainties are actually quite rare, and there will always be an amount of unknown or unknowable information that will either require the organisation to apply contingency measures (to account for the uncertainty) or simply to take a gamble (to simply find out whether something will work or not). It is important to understand that there is not necessarily anything wrong with this, but there is a “wrong way” to do it. For example, the failure to account for fairly likely and foreseeable events in your contingencies and simply “hope” they don’t happen, or the failure to apply bearable limits to your gamble – understand the worst case scenario and make sure you can live with it if the worst does happen.