John Seddon is a visiting professor at Cardiff University, and managing director of Vanguard Consulting. A frequent speaker at conferences, seminars, universities and business functions, John has a reputation for being outspoken, controversial and insightful.
Even though he is a student and vociferous proponent of Deming and Ohno, John’s views frequently clash with those of Six Sigma purists. He believes that too many practitioners place too great an emphasis on the tools instead of the work, and promotes a combination of Systems Thinking and intervention theory as a way of implementing successful change programmes in organisations.
John is the best-selling author of “Freedom from command and control: a better way to make the work work”. His latest book, “Systems thinking in the public sector: the failure of the reform regime and a manifesto for change”, is available from Triarchy Press.
_______________________________________________________
improvementandinnovation.com: You trained as an occupational psychologist. How did you take that into the business world?
John Seddon: I was a clinical psychologist at first, working at prisons trying to help them and I met a man who had the chair at London University in occupational psychology and he said to me: “all the effort you expend trying to help individuals, you’d be better off working on the organisation because you’d have greater leverage.” So I thought about it for a few years and began reading about organisations and the design of organisations, I went to study under him. He was a big influence on me, his name was Nick Giordiades, and with him I did an MSc in occupational psychology.
One of the most important things he taught me was something called intervention theory, which is how do you go into a system full of people and intervene in it in such a way that you leave it better off and capable of solving similar problems subsequently.
Eventually I became a consultant, working at the big change programme at British Airways in the ‘80’s. It was a very important programme [but] I still had nagging doubts because I was wondering that although the programme was clearly working, could it be better? Even though we’re going in and changing organisations, do we really know how they work?
It’s an odd question. And it was around that time I hit Deming, and he was absolutely brilliant on explaining just what was wrong with management. When I read “Out of the Crisis”, I just couldn’t put it down. I could see what he was talking about. He would say that this is manufacturing and you need to see the whole thing as a system; but how do you do that for service? So I set upon that problem, and over the years we started to get the idea of how to do that.
I read a lot of systems thinkers, like Peter Senge, who was interesting and illuminating, but didn’t really help in terms of what you do on Monday. So I had to kind of work it all out myself. In the middle of working it out – and we’d gone a long way towards solving some of these problems in how to see things as a system – I came across Taiichi Ohno’s work. And of course he built the system designed to build cars at the rate of demand. I could see that there were some principles in there that were important like design should be in process, like you’ve got to manage flow and so on. But what you do in service is different to what you do in manufacturing because they’re different beasts.
For example, you’re not solving the problem of how do you move all the physical components that you’re manufacturing into an assembly at the same rate, which is what you’ve got to do if you’re managing the flow of materials at the rate of customer demand. So all of the tools associated with managing flow in manufacturing are for solving that problem. In service you’ve got a different problem.
Each of these events gave me insights, but not answers and in fact more questions. So myself and a very small gang – there was about six of us in those days – made it up, worked it out. It’s still new.
How do you define a system?
I put a boundary with the customer. The important thing about a system is that the parts don’t act independently: they are interdependent. You need to understand the interdependencies in order to optimize – Deming’s word – what you’re doing. So departments are interdependent, but also features are interdependent, so the way a process works will be governed a lot by the way it’s measured.
In the public sector for example when you put targets – arbitrary measures – into your process you really mess your process up. So it might look like you’re doing great, but you’re not. Interdependencies are not just functional, they’re abstract things that come together.
You talk in your book about how the different departments of the public sector effect each other: would you see the public sector as a system as a whole?
You first learn that when [goverment departments] don’t actually deliver good service, then you get what I call “failure demand” arriving at local authorities, voluntary agencies, legal services, etc. We’re talking very high volumes here so you need to look at that together, because the only way to cut out the billions that we’re spending on downstream, is to re-design the service.
Is failure demand the same as NVA?
No. Failure demand is a phenomenon that is peculiar to service organisations. There are other demand phenomena in manufacturing, like demand amplification, which you see in the beer game, as in Senge’s example. False demand signals things like only being able to buy one colour of car. But you don’t get failure demand, and failure demand I defined as demand caused by the failure to do something right for the customer. “I don’t understand my bill”, “it hasn’t arrived”, “you haven’t fixed it”; this is all failure demand. You do get a bit of it in manufacturing but it’s not as big a concern, and it’s not central to the design of manufacturing operations the way it is central to the design of service operations.
So failure demand is a major type of waste; in fact it is always the largest type of waste in service operations. We discovered this when we were working for a guy selling computers over the phone – we’re talking here about 1985-86, it was a very new concept as PCs had just arrived. He had a productivity consultant working for him as he wanted to increase his revenue and the consultant suggested giving operatives an incentive: anything they sell over £100 they get a bonus of £5.
This is an inbound call-centre, and I thought to myself: ”there’s a risk here.” Someone might ring up for something today for a fiver, but they might phone up tomorrow for something for a thousand pounds. We don’t want to give them bad service but where’s the incentive? Maybe we should find out more about why they call, because it’s inbound, all the operatives have to do is pick up the phone. And then we had the most amazing shock because we spent about three days studying demand, and we discovered something that 45% of the demand hitting the system is what I now call failure demand.
So I turned round to him and I said: “if we want to make more money here the answer is deliver this service in such a way that the customers don’t ring up [to complain] and that will increase our capacity, we’ll take more calls and we’ll make more money!” That’s a system idea, whereas a command and control thinker would say: the way to make more money is to lean on your people.
Womack and Jones invented the word ‘Lean’ because managers want things in a box and therefore you need to give it a label. Ohno said to never give it a name, don’t codify the method. It’s how you think that is the key, how you conceptualise the problem. If you were learning from Ohno, he’d put you in the factory, tell you to study the factory, and then he’d go away for two weeks. Then he’d come back and ask what you’d learned. And if you hadn’t studied the things you want to learn, he’d show you what to study, and then he’d go away again.
Some people would say: why doesn’t he just tell them? It’s because he knows what he’s got to teach you is how to work it out. He’s working on your thinking, not giving you the tools.
Lean is focused on the customer and a lot of what you are saying is the same.
A client of ours who undertook a Lean programme saw no major benefit to the bottom line because all the reported improvements were related to activity. The Lean ‘toolheads’ had bought into the command and control idea that work is simply activity, and activity equals cost. Therefore if we go in and look at a bit of a process and cut down some activity, so we reduce the processing time. That reduction activity will translate through to real savings. Wrong. Cost is in flow, it’s not in activity. By tinkering with a bit of the flow, you’ve got no idea whether you’ve had a positive or negative impact on the overall flow.
But in reducing the time, surely you are reducing costs at some level?
That’s what they think, but it’s wrong. Manage costs and your costs go up, always. For example, what we do is we go and have a look at an intervention where there are lots of reported improvements. Let’s take something like new business in pensions processing. What matters to a customer here? End-to-end times are usually a pretty good measure: from the time I say I want to become a customer to the time I’ve actually got the product. That’s the first thing a systems thinker would measure to get an overall picture, before you go down. A Lean thinker wouldn’t measure that, they go straight in with “what’s your problem?”. And they are told something like “this operative here does eight things an hour, whereas this operative here does 16 things an hour, and if I can get all of them doing 16 things an hour then….”
The first problem here is: is eight the same as 16? We can measure that and what you usually find is that they are the same. In other words, they’re just as probable; the variation of the work has got nothing to do with the workers, it’s got everything to do with the system. You must therefore study the system. People make the mistake of assuming that they have to get productivity up to the performance of the best and they think that’s the right problem, but usually the problem is completely different to the one that you think you’ve got.
So, going back to managing activity, this is why they [outsource] to India. Because they think that work is all about cutting up work into sub-tasks, sending the easy tasks off to India or wherever, because you get them done cheaper, and then you get them sent back. Well, what you learn when you study the flow, is that very often, outsourcing work to India simply increases the amount of demand on your system. So you’ve got one cheaper call in Delhi driving up the total calls you’re taking for the customer to get a service, so your costs go up. Do you suppose anybody measures cost that way?
So you’re saying that by saving money through outsourcing, costs may have gone up elsewhere?
That’s correct. You have to get the big numbers to start understanding the big relationships and then you’ve got to study the work to understand the relationship between what’s actually going on in the work and these big numbers. And it’s a really salutary experience for the leaders.
It also raises another issue – and this comes with intervention theory – who should do that? A consultancy firm might send in 25 suits and they’d do that analysis but my view is that the executive team should do that work. My purpose is making sure they do it right, because they’re going to get some major, counter-intuitive moments.
If you’re an interventionist and you read Ohno’s work you realise that he discovered counter-intuitive truth. For example, you might believe that activity equals cost. I know that isn’t true but if I try to dissuade you of that idea, you will either go away disagreeing with me, or kind of curious about it. However, if I give you work to do where you discover the truth of that idea, there is no argument.
How does getting rid of activities that create waste make things worse?
Because you don’t know if your improvement in this little bit is an improvement end-to-end. So you get these executives who go and study what they did, and you often find that while the activities have been improved, the end-to-end time for the customer got longer. We’ve got to understand the total cost of the system, not get into managing micro costs.
I teach managers how to manage value. They have no map of how to manage that, so you have to teach them and you do that at their work, not in a class room.
You have been critical of Six Sigma in the past. What do you see as being the main problems with it?
Six Sigma was invented by Motorola, based upon the work of Taguchi. Whenever you come across a tool, you need to ask: who invented this? What problem were they solving? Do I have the same problem? In Motorola they were working on Taguchi’s idea that reduction in variation will increase your quality. Because it worked well, they wanted to know how they could apply it elsewhere, which is an interventionist question, but they came up with a command and control answer. Essentially, this was to train people, for them to have projects, and for them to report on these projects.
The problem is that they don’t know what their problems are, or they’re trying to solve the wrong problems. They’re trying to solve problems like “how do I get my engineer to answer more calls in a day?” Wrong problem to solve. If you solve the system problems, he’ll do more calls.
What about the companies who have had a very successful Lean Six Sigma programmes?
There’s a lot of rationalization in management. They think that these programmes are helping on the general journey and sometimes there are true improvements, because despite all the trappings there may have been some talented people in there that understand how to do good stuff.
It goes back to if you are a manufacturer and you want to use Taguchi’s ideas to reduce the variation in components to improve your quality. That’s all well and good, but how do you get others to do it? Teach them how to think. Teach them how to look. And then they’ll pull the tool. Six Sigma interventions are a push: managers define the problems and push the tools.
You said earlier that redesign is not top-down, but outside-in. Is that going back to starting with the customer and working in? What do the top people need to do?
They need to understand the philosophy, and be able to work on the work with the workers. I think one of the important things about the leadership literature – and therefore leadership interventions – is that the leadership world has got disconnected from the organisational context. We talk about leaders having visions and missions and values, but it’s all tosh. Ohno would say to his managers “if you’re not washing your hands three times a day you’re probably in the wrong place.”
How do you define leadership?
Working on the work with the worker. Prosaic, eh? No visions, no values.
Surely people need to know where the company is going and what the plan is?
Those are command and control ideas. You said earlier ‘it starts with the customer’. Yes, but, the most important thing is it starts by understanding demand. That means you don’t go and do focus groups with customers. If you study demand it will be loud and clear what’s going on from the customer’s point of view. Strategically, you design a system that is adaptable to customer demand, so as demand changes your system changes. How cool is that?
So now you’ve got an adaptive system which is a strategic opportunity because the customers will pull you away from the competition. The customers will also tell you about services and products you should be providing and so on.
What if you have 45-60% failure demand?
You have to design the failure demand out. There was a book published recently called The Best Service Is No Service which takes the failure demand idea – although it doesn’t call it that – and they propose that if you get your service to work, customers don’t ring up, therefore it’s cheaper. They’re right, but why express it as “the best service is no service”? From a customer’s point of view the best service is a service that works for me. In Taguchi-speak the customer sets the nominal value.
The reason they call it “the best service is no service” is because they are appealing to a cost audience. Wrong idea. You cannot turn off failure demand without re-designing your system, and you’re going to find that the principal causes of failure demand are your current measurements: rules, procedures and structures.
Another thing that’s important is that you can only remove failure demand that’s predictable. So before you set about changing your system, you have to establish which of these failure demands are predictable and which are unpredictable.
This is something Deming taught; it’s all part of the theory of variation. Things will always go wrong; you need to understand what’s going wrong predictably. Things that are going wrong predictably are under your control, so you re-design. Things that are going wrong unpredictably: don’t touch them. Because if you treat the unpredictable as though it is predictable, you’ll add to the complexity, which is what managers do all the time because they don’t understand the theory of variation
What question are you always asked?
“Is this a leap of faith, John?” They feel like it’s a leap of faith because they hear all these things that are unfamiliar. But what they have ignored is that the first major step is to study the current work as a system, the object of which is getting knowledge. Of course, knowledge is the other side of risk therefore there is no risk, there is no debate, there is no leap of faith. Knowledge. It’s a rare commodity amongst managers.
You can’t re-design until you’ve got that knowledge. So understanding which demand is predictable is an example of the kind of knowledge you should get, and once you’ve got that knowledge you can then move into re-design. So it’s not a leap of faith. If you find when you study it as a system that there’s no scope for improvement then you’ve learnt you’re fine. But if it is a command and control organisation then it will be stuffed full of waste. They always are.
