Soft Systems Methodology: Part One

Robert N. Phillips
CEO, Lasotell Pty Ltd.

How many times have you tried to “improve” a “system” in the workplace only to find that, after all the hard work and after deep-down, middle-of-the-night, honest reflection, the improvement has been only a modest success? Are you still basically amazed at the number of things that did not seem to fit quite the way you thought they would? Do you also feel, in all honesty, that the longevity of the new system has more to do with your continued presence rather than with people believing in its inherent worthiness?

Why do such things happen? After all, you identified the requirements, planned the work, and did all those good things. Why did the “owner” of the system and everybody else keep finding all these wrinkly, irritating bits and pieces along the way?

This article is the first of several articles about Soft System Methodology, which is best described as an organised way of tackling messy situations in the real world. Some examples of messy situations are

  • improving health care for elderly in your local district
  • deciding how the company should take advantage of information technology
  • improving the productivity in a department
  • creating a new workflow system for three teams of people
  • planning your career
  • running a sports club

(These articles will draw heavily on the book Soft Systems Methodology in Action

[Wiley 2000].)

Soft Systems Methodology has its beginnings in the 1960s when Gwilyn Jenkins became the Professor of Systems at Lancaster University. His appointment was the first of its type in Great Britain. Jenkins’s aim was to “find ways of understanding and coping with the perplexing difficulties of taking action, both individually and in groups to “improve” situations which in day-to-day life continuously create and continuously change.” The intervening 30 or so years have been a process of exploration and evolution of thinking to get to the Soft Systems Methodology of today. Peter Checkland has been a dominant figure in this work for the last 20 years.

So, let us begin.

We all tend to look at a problem situation and say “all we need to do is….” But a set of circumstances that can be solved that easily is, by definition, not a “messy situation.” Nevertheless, many people try to address a wide range of problems in this simplistic way. So, why does it not work? Two reasons. The easy reason is that “messy situations” are multi-factorial. There are always a large number of factors driving the situation, which may or may not be linked horizontally or vertically. And of course, there are all the factors on the next layer or two up from the situation we are trying to resolve. The second reason is the hard one—messy situations always involve people.

The first approach to addressing messy situations is to try applying pure System Engineering principles. After all, that approach works well when the objectives are clear and the requirements are universally agreed upon. But in reality, people have their own views and judgements and their own standards and values. So trying to define an agreed set of objectives or requirements in a messy situation is like trying to corral cats! If there is any doubt about that, consider the political din that arises when anyone suggests changing the Health or Education system at a state or federal level, and think about how long it takes to actually make any change at all.

One of the fundamental aspects of the Soft Systems Methodology is understanding how human beings interpret the world around them. Basically, we build a mental model of how we think a device or human “system” operates and then we take actions based on that internal model. As experience teaches us more about the device or system, we modify our internal model. When reading Checkland’s description of that process, one cannot help but be reminded of the highly illustrative example given by Donald Norman in The Design of Everyday Things (Currency/Doubleday 1990) showing how the average person construed an internal model of how the cooling system worked in a particular refrigerator that was poorly designed. (Norman’s book is still available at and we will discuss it further in this column in the next issue.)

In principle, the human process of model building is a good approach, but it has several flaws. There is a limit to the complexity of the model one person can create and retain in the mind. But more important, how accurate and how complete is the model? If the model is deficient, the outcome will also be adversely affected. This discussion brings us to the need to have a way of gauging the appropriateness of the mental model.

We will commence the next article with this topic.


Soft Systems Methodology in Action
Peter Checkland and Jim Scholes
2000, New York, NY
John Wiley & Sons
ISBN: 0471986054

The Design of Everyday Things
Donald A. Norman
1990, New York, NY
ISBN: 0385267746