I don’t think this is an actual Dalek quote, but it should be….. “Failure is not an option…. exterminate!”
I’m back to pursuing learning from failure after listening to an episode of The Life Scientific on BBC Radio 4, featuring Professor Dame Wendy Hall. In response to a question (27mins) on why the UK doesn’t have the equivalent of a Google or Twitter Professor Hall responded; “we don’t kill things fast enough…. most things will fail…. “. This was a thought-provoking answer and is worth listening to. In fact, listen to the whole programme if you are at all interested in computing and the discipline of Web Science, which she co-founded at Southampton University.
From this explanation of failure in the complex world of technology innovation I ran straight into one of my Dalek friends….. a practitioner from the world of Heath and Safety regulation.
Whilst we were chatting about my recent blog posts on failure they did actually say “In my line of work failure is not an option…. things are either right or wrong. There is no place for failing and learning. We have to be right” (best spoken through one of those voice changer toys….. EXTERMINATE!).
Well they are both right. In some cases ‘most things will fail’ and we need to ‘kill things quickly’, learn the lessons and move on. In other cases failure isn’t an option, things have to be right. The response to failure is very specific to the situation. I’ve been trying to find a good explanation for why this is the case, and here are two examples; one by John Caddell who wrote The Mistake Bank and The Cynefin Framework by Professor Dave Snowden.
John Caddell, The Mistake Bank.
John Caddell has been very generous and shared material with me on how to learn more effectively from failure. One of the key points I took from this was the need to understand your circumstances. How you relate to failure in a highly structured, predictable situation will be different to a highly unstructured unpredictable environment. This diagram from John’s Mistake Bank blog has been adapted from the work of Professor Amy Edmondson. Professor Edmondson wrote, Strategies Learning from Failure, which she discusses in this Havard Business Review video.
The key point is, for highly predictable, less complex areas, failures should be treated as defects and avoided. My Dalek friend is right here. A 3 ton load in a 2 ton limit should not be allowed. In unpredictable and highly complex areas, failure needs to be looked upon as something you can learn from. The trick here is to develop an approach that allows rapid learning through fast intelligent failure. I find the Cynefin Framework helpful in understanding this.
The Cynefin Framework and learning from failure.
The Cynefin framework is also the subject of a Harvard Business Review article so has a good pedigree.
From the perspective of learning from failure it is worth focusing on the four domains. In the simple domain there is best practice, one way of doing things. Here failure, deviation from what is ‘best’, is a problem. Rather than accept failure and learn, its best to avoid the failure in the first place, as there is little to improve.
Jumping straight to the complex domain. There is no known answer, practices are emerging from a complex unknown situation. This very much describes the situation in the modern world and the example given by Professor Wendy Hall. In technology innovation there are lots of ‘attempts’ and failures need to be ‘killed off’ quickly. The crucial point is to learn from the mistakes and move on. In the words of the Cynefin Framework, ‘probe’ for new solutions. More on the Cynefin Framework can be found in this video on the Cognitive Edge website.
Daleks, the enforcers of Best Practice. I have limited experience of real Daleks but I do think they would be enthusiastic ‘best practice’ enforcers. Insisting on the one and only ‘best’ method of doing things. This would be regardless of the level of uncertainty or complexity of the situation. Failure is not an option……..EXTERMINATE!
By contrast, John Cadell, and three Professors all say that there are circumstances where failure in inevitable and should be encouraged so that you can learn, make changes and move forward. The world of emerging practice, not best practice or even good practice.
Failure is an option, it just depends upon the circumstances.
So, What’s the PONT?
- Learning from failure is a highly valuable activity that should be encouraged.
- Understanding the situation and what approach to apply to failure is important.
- Daleks and Doctors (Professors) have their rightful places, but mixing them might not get a happy ending (YES! A Doctor Who link at last)
Picture source: BBC Doctor Who Publicity. http://www.bbcshop.com/doctor-who/doctor-who-talking-dalek-bronze-dalek/invt/04352a
Chris, your blog got me thinking about when things do go wrong, they don’t totally go wrong, there is always some learning. Like Pareto’s rule, whilst possibly 80% of something might be wrong, 20% is right. Taking the learning forward, means lessons can be learnt. Acknowledging this is important, because as a society, we Brits are so very black and white, like the health and safety friend you mention in your blog. The reality is, grey is the predominant colour.
It’s worth looking at the Prof Amy Edmondson HBR video.
I agree that there is learning that can be extracted from all failures.
The big question is around the approach.
She uses the example of high volume precision engineering in automobile manufacture as an example where you need to get it right every time (same would apply to aircraft maintenance).
The tolerance of failure is very low in these situations.
If you need to fail and learn, do it in a controlled test environment.
For a developing area full of unknowns and no right answered, the fail and learn approach should be encouraged. However, there also needs to be some structure and control here, to make sure the learning is extracted and to quote Prof Wendy Hall, you kill things early – the ones that don’t work.
Thanks for the comment
All very interesting stuff and John Caddell and Amy Edmondson are great sources of information on failure. Undoubtedly failure needs to be looked at differently depending on the complexity involved. In Strategies for Learning from Failure, Amy Edmondson provides us with a spectrum of reasons for failure, ranging from blameworthy to praiseworthy and related to the nature of the work involved be it simple, complex or experimental, with failure in experimental, exploratory work in certain instances being praiseworthy. What interests me more in a way though is the general attitude to failure related to overall performance, whether that be personal, project based or organisational.
I recently discovered this simple but great quote from Dr. Lucian Leape MD, a physician and professor at Harvard School of Public Health
“The single greatest impediment to preventing harm is that we judge and punish people for making mistakes.”
He was discussing preventing harm to patients.
I find that most organisations as with most people are reactive to failure rather than proactive to it. In other words before we can learn from failure we have to actively seek it out and we need to encourage the reporting of it.
In the Swiss Cheese model which I think you are aware of, small failures eventually line up to produce abject or catastrophic failure. So even the seemingly trivial may need to be looked for and the finding and reporting will probably not happen in a judgmental and punishment based culture.
Amy Edmondson in the same article talks about it in terms of creating a psychologically safe environment.
Failure in my view is a constant, it’s always there somewhere. We don’t have the power to allow or disallow it, only the power to interact with its presence in a way that is productive or non productive.
Making the same mistake repetitively in a simple or complex operation could be said to be blameworthy but maybe our default position on failure should be that the only blameworthy act in relation to it is not reporting it.
Thank you Roxanne and Martin,
There is a lot of good food for thought in this comment.
I am particularly taken by your view that failure is a constant and we don’t have the power to allow or disallow it. The only power you have is how you respond.
I think this is particularly relevant in many of the large and complex orgaisations that deliver public services. In an orgainisaton of seeral thousand people, delivering dozens of different services to many and varied individuals, anyone who thought there wasn’t something going wron is fooling themselves.
Coming back to how you respond to failure is so important. I do wonder if there a degree of social pressure to always treat frailure as a bad thing. A while back Bryn Williams and Matt Wyatt had a discussion in the blog comments around the statement that “as a species we are anthroplogically conditioned to avoid failure rather than accept good practice”. There was a fair bit of the discussion that was around the anthropiological aspects of failure (I’m deliberately avoiding the evolution word here – got me into deep water last time).
The comments from Bryn and Matt can be found here: http://whatsthepont.com/2013/08/06/are-we-programmed-to-innovate-or-stick-with-what-we-know-welcome-to-the-jungle-and-the-big-beasts/#comments
I have a colleague with background in anthropology. I think I will ask him if he knows of any societies where failure is treated in a way that isn’t wholly negtive….. I’m quite intrigued by the prospect.
One thing has just struck me the though….. The Eddison story about the 999 failed light bulbs and success on number 1000 is quite well embedded in coprorate story telling. I wonder where this story sits alongside those of disasterous (avoid at all costs) failure? It only seems to be relatively few stories about failure that are associated with ultimate sucess that get told in comparision the the warning tales?
Lots more to say,
Probably best kept for another post, or when we eventually get to meet.
Thanks for taking the time and sharing your thoughts.
Greetings from Madrid. I’m guessing I’m the anthropological colleague you’re going to ask. The short answer is no – I can’t recall any societies where failure is treated as other than negative (though I haven’t actually looked at that many!). However, you do raise some interesting questions in my head.
In general, the small hunter-gatherer societies that anthropologists typically study tend to be quite conservative and have pretty rigid social and cultural mechanisms through which change is mediated. Conformity rather than innovation tends to be valued, though it isn’t clear cut in practice.
When it comes to failure, probably the biggest failure they face is the failure of the hunt. Here I generalise, but the tendency is probably to blame such failures on the supernatural rather than individuals (though individuals may get singled out if sorcery is considered to be the cause).
My view (possibly shared by others, it isn’t something I’ve looked into!) is that this probably serves the purpose of socialising failure. Rather than being down to the fault of one individual, responsibility for failure comes down to the whole society not properly worshiping the ancestors/ breaking taboos/ failure to follow tradition etc. There is a good example on hunting failure amount the Inuit at the start of this paper http://lycadican.sourceforge.net/papers/rel_anthro_paper.html . Whether you consider using failure as an opportunity to reinforce tradition/ social bonds is a positive response is a matter of interpretation.
At face value you could say the lesson is that there may be something intrinsic in human nature that tells us we need to conform/ do things the same way and failure reinforces the correctness of the old ways. But in fact, those societies aren’t entirely static and do innovate. They do subtly change hunting habits – many use modern weapons – and if they didn’t adapt and innovate there would only ever have been one method of hunting. But those changes probably reflect innovations that worked – so the hunt didn’t fail and there was no appeal to the spirits/ ancestors to explain and forgive the breach of tradition. And I suspect many innovations were woven into existing mythology/ tradition or revealed through divination/ shamanism to get the legitimacy to be used in the first place.
I’d need to give some further thought to what the real lessons may be. I think there probably are some useful insights from anthropology on socialising failure through organisational culture. I suspect there is an unavoidable lesson that innovations that don’t fail are least likely to bring about the wrath of the gods/ peers! Also, when I eventually get to my blog I plan to write something about innovation and culture change and how the new is mediated through the language/ symbols/ organisational myths of the old.
A couple of caveats – I’m probably doing too much generalising and I’m not sure to what extent these ideas are actually reflected in the ethnographic evidence. Also, I’m writing after a few glasses of a fine Ribera del Duero accompanied by too much jamon iberico de bellota!
Hope this makes sense and provokes some thoughts!
Thank you Mark.
It’s getting late here so I’m having another good read in the morning.
I know someone else who will be fascinated by what you’ve written, apart from me.
Hope you are enjoying Madrid, and the local delicacies.
I interpret Professor Hall’s comments as meaning simply poor project management.
I think that your Dalek friend is well equipped to do their job – but realises if they were creating a competitor to Google failure might be better known to them.
I too come from a place where failure was not an option, but we failed all the time!
However we were good project managers, and while elements failed every day we always got back on track and the project end date was always achieved.
I worked for Honda manufacturing – they would make grand statements like
“On Tuesday the 18th of September at 2:15 in 4 years time the first new Civic will roll off the line” They would make this type of statement about everything they did. Even new car plants and production lines.
I don’t know about the Mistake Bank or the Cynefin Framework – but I do know that anything like this that I have ever encountered since leaving Honda has simply been Kaizen. (Re-marketed).
The P,D,C,A Cycle – Plan, Do, Check, Action
Or the Cap Do cycle – same thing.
Plan it – Do it – Check it (measure the results?) – Action (Asses what’s going on – propose improvement)
Then you repeat the cycle it’s no use without constant attention and measurement
Perhaps we weren’t constantly failing, just pursuing best practice.
There are other names for it including the Deming Cycle, but I believe even Mr Deming borrowed it from the world of science.
As for assessing what was going on and proposing improvement. Or Situation Analysis.
There are 2 steps to take
1. The 3 reality principle – Go to the actual place, understand the actual problem, be realistic.
2. List every element of the situation or problem on a table. to determine root cause and list of possible actions. Background, Problem, Cause, Countermeasure
I strongly agree that we should learn from failure, and in order to recover from failure we must understand it’s cause so that we can plan our way out of it. We must then manage our way out of failure never taking our eyes off the things we measure.
Should also mention “There can be no improvement without measurement”
So why haven’t we (the UK) created a twitter or a Google?
Could be poor management, that’s true
I think it’s a numbers game there have been millions of endeavors on the internet, Facebook, Twitter and Google are just three of them. America is a big country and we are a small one.
I listend to that section of Prof Hall’s interview a couple of times and got the impression that it was more to do with financial backers not wanting to pull out and loose money on an investment. I have heard this mentioned a few times.
The nature of developing tech ventures from concept to market place is very high risk and unpredictable. The quote from Prof Hall that ‘most things will fail’ seems to be widely accepted.
In a way this seems to fit with your numbers game comment. There are millions of endevours, only a few suceed. Prof Hall seemed to be saying that in the ‘old world’ we are more inclined to hang on much longer the see if things might work.
It is an interesting area.
I think the ‘experimental’ nature of what is being done leads to a different approach to failure that might not be present in a more standardised. high volume activity.
Interesting what you say about Honda. I included this video in a recent post about Honda’s approach to failure.
To be honest, I did wonder how the approach described in the video would fit in with a high volume, high quality precision engineering activity. The notion of failures being welcomed on the full scale production line didn’t seem to quite fit with the world of LEAN and Six Sigma.
There is a lot of PDSA /PDCA cylce present in all of these approaches, I suppose thats a measure of just how useful it is, likewise Kaizen.
Would love to hear what you think of the video.
Thanks for the comments,
Must confess I didn’t listen to the woman in the earlier interview and what you say does resonate and makes a lot of sense.
I’m very glad that you’re up on kaizen.
I really want to jump that and get onto the Honda video ;0)
Where do I start?
I was there in ’94 ’95 ’96 when in a 3 year period we did very well with our F1 and Indie car engines.
Honda is complex.
If Honda had a bad week in F1 they would rebuild the engine and win the race the following weekend – there must have been significant risk of failure in this approach – very aggressive!
Honda’s background is in racing.
I worked in purchasing and asked my Japanese boss “Did you know Mr Honda?”
He said “Yes, but he never hit me!” It turns out that Mr Honda hated to be told ‘no’ by us bean counters and was an engineer at heart, he would attend meetings with spanners in his pockets. He was a brilliant engineer, and didn’t allow anything to interfere with his ideas – especially small thing like money or the risk of failure.
Didn’t Henry Ford say “The man who never made a mistake never made anything”…? Well it could easily have been Mr Honda.
On the production line everyone knew it’s ok to make a mistake as long as you put your hand up so that we can fix it before it gets into the hands of the customers.
Given the flirtatious attitude towards risk in the video – I have a few things to say about that.
It always seemed to me that the further away you got from manufacturing the more stupid people were – eg the boss of the woman who wanted to have orange cars!
I bought the heater and air-con for the Accord – my supplier wanted to change the plastic material used to make the heater case (French company) anyway plastic prices were going through the roof and they had a cheaper and almost identical alternative. Honda refused the new material and would not allow the change until the next minor model change. This would also allow time for very severe testing of a new material not known to Honda.
My supplier had made running changes to the other manufacturers he supplied and they all accepted his word that the new material was good enough. Armed with pages of technical data and samples I went to our engineering department to try and push the change through. I said “everyone has this new material Saab, Volvo, Renault, Peugeot etc” The engineers said “We are NOT Saab, Volvo, Renault, Peugeot etc” I got by back side kicked – we made no change!
Our ABS braking system was not like any other – it was over engineered as everything Honda did was. We would not adopt external technology until it had been in the field and proven for several years. Regardless of cost.
Honda will not allow the customer to be subject to any risk.
But, they do push the limits and their “What if?” campaign sums them up.
Look at Asimo – he is 15 years old. Honda is not a car company they are an engineering company.
Interesting that the video is aimed at the American market.
Yes without risking failure you will never break those boundaries, and racing has traditionally been their proving ground.
Look at the Integra Type R straight off the production line and onto every race track in Europe where it smashed every lap time for a 1.8 litre racing car ever set, and straight into the showroom. Why? Because they can!
So to answer you question about mixing risk and volume production.
Honda are the parent who hits the drinks cabinet, but always locks it afterwards.