Systems

Diagrams

Env
3

iron
2
2

1 1 men
t
1 1

1 1

System 1…operative
unit
System 2…coordinatio s
n and information

A practical guide
System 3…manag
ement and control
System 4…intellig
ence, monitoring
System 5…policy, of environment
coordinates internal
and external, present and future

BobSystem
Williams
diagrams
 SYSTEM
DIAG AMS
A P ACTICAL GUIDE

Copyright © 2021 by Bob Williams

The moral rights of the author have been asserted.
For Ariel
ISBN 978-0-473-58457-7 (PDF)
ISBN 978-0-473-58457-3 (EPUB)
https://gum.co/systemdiagrams

You are welcome to quote from and copy parts of this workbook for
your own use according to the usual Creative Commons principles.
Please quote the full source when you do so. And please let me
know about it; I’m curious to know the nooks and crannies this ends
up visiting.

Enquiries concerning commercial reproduction rights should be

addressed to the author at bob@bobwilliams.co.nz.

All URLs worked July 2021

System diagrams
R
R
Acknowledgements
My thanks to all those who have helped in the production of this
book. In particular I would like to thank those who read and
commented so enthusiastically and valuably on the draft book:
Steve Powell, Cristina Planchart, Beatrice Tschinkel, Jan
Brouwers, Deva Henry, Silva Ferretti, Frank Page, Steve
DeLorey, Julie Witmer, Leigh Dowsett, Bernard Crenn, John
Hoven, David McClintock, Mark Anderson & Dirk Frans.
Thanks as always to Martin Reynolds for his unwavering support,
whip-smart critiques and invaluable advice, Judy Oakden for her
generosity and collegiality, Patrick Hoverstadt and Terrence Love for
helping me unravel the subtleties of VSM. Sue Carswell and Diana
Beattie for their diagrams.
My thanks to Barbara Schmidt-Abbey and Joan O'Donnell for many
things but in particular allowing me to (ab)use their research when
creating the VSM example.
Deeply indebted as always to my long su ering copy editor, Nikki
Crutchley of Crucial Corrections. nikki@crucialcorrections.co.nz
Katrina Mitchell of Picture Impact waved her magic wand and
transformed the look of this book into something you might actually
want to read. katrina@seebeautifuleveryday.com
Finally thanks to everyone who has written about or spoken on
system practice and systems diagrams. I owe a huge debt to you all.
Bob Williams.
Wellington, Aotearoa New Zealand
http://www.bobwilliams.co.nz

August 2021

System diagrams
ff
 Table of Contents
Why Systems? 13
The systems eld ..............................................................................................13

What is a system? ..............................................................................................14

So what is a systemic approach? ..................................................................16

Introducing System Diagrams: Why, What, When, Who

and How 18
Bob Williams 1
Is it a depiction of a system? ..........................................................................18

Why a Book on System Diagrams? 2 Two distinctly di erent kinds of system diagrams ...................................19

About system diagrams ....................................................................................2 Selecting the boundaries of system diagrams ..........................................21

What’s in the book ..............................................................................................3 How to decide key boundaries ....................................................................22

Facilitation ............................................................................................................3 What questions should diagrams pose? ....................................................24

More about this book ........................................................................................4 How should diagrams be interpreted?.......................................................25

How to read this book .......................................................................................4 Now what? What next? ...................................................................................26

Further resources ...............................................................................................4 How to identify ‘Now What’ and ‘What Next’............................................26

Why Diagrams? 5 CONSTRUCTED DIAGRAMS 28

Why would you want to diagram something or interpret a diagram? ..5 Resources ..........................................................................................................28

What comes to mind when you read the word ‘diagram’? ......................6
Rich Pictures 29
Di erent realities, perspectives, framings and purposes ........................8
Purposes of Rich Pictures ..............................................................................29
The purpose of the diagram ............................................................................9
What does a Rich Picture contain and what does it look like? ............30
The meaning that people draw from the diagram ...................................10
How to draw a Rich Picture ...........................................................................32
So what for diagrams? .....................................................................................12

System diagrams
ff
fi
ff
 In uence Diagrams 36 Cyne n questions.............................................................................................74

What is an In uence Diagram? .....................................................................36 Cyne n resources ............................................................................................74

Why draw an In uence Diagram and not a Rich Picture? .....................38 Viable System Model (VSM) 75
In uence not causality ....................................................................................39 The purposes of the Viable System Model ...............................................75
How to construct an In uence Diagram ....................................................40 The VSM diagram explained .........................................................................75

Causal Diagrams 46 The VSM analytical process ..........................................................................77

Purposes of Causal Diagrams ......................................................................46 VSM as a re ective process ..........................................................................83

Causal Diagrams and In uence Diagrams ................................................46 Recursiveness ...................................................................................................84

Multiple Cause Diagrams and Causal Loop Diagrams ...........................47 Viable System Model (VSM) resources ......................................................84

The features and conventions of Causal Diagrams ................................49 Cultural-Historical Activity Theory (CHAT) 85
Framing the Causal Diagram ........................................................................54 Purposes of CHAT ...........................................................................................86
Analysing Causal Diagrams ..........................................................................55 CHAT history .....................................................................................................86
How to construct Causal Diagrams.............................................................55 What is CHAT? ..................................................................................................86
Questions that help analyse and re ect on Causal Diagrams .............63 Constructing the CHAT diagram ..................................................................87
A Coda: Causal Loop Diagrams and System Dynamic Modelling ......64 CHAT resources ...............................................................................................97

GENERIC DIAGRAMS 65 So Where Do You Go From Here? 98

Cyne n 66 Some advice from the trenches ...................................................................98

Purposes of Cyne n ........................................................................................66 Think systemically, be systemic ...................................................................98

The ve Domains described .........................................................................67

Cyne n as a tool for managing situations .................................................69

System diagrams
fl
fl
fi
fi
fi
fi
fi
fl
fl
fl
fi
fl
fl
fl
 Bob Williams
Other Books and Videos
History Systemic Evaluation Design. 2nd Edition (2016)
https://gum.co/evaldesign
I am known internationally as working in various elds, including Video here https://youtu.be/VUTqG5j21mU
systems, evaluation and action research.
Uso de Conceptos Sistemas en el Diseño de Evaluación. – Un
I have been using systems concepts in my work for over forty years. I
Documento de Trabajo (2017)
was originally trained as an ecologist, one of the earliest systems
https://www.evalperu.org/wp-content/uploads/2021/04/Libro-Bob-
disciplines, and spent four years with the Systems Group at the
Williams.pdf
Open University in the United Kingdom. I am well versed in a variety
of di erent systems methods, although in recent years have focused
Wicked Solutions: A systems approach to complex problems. 2nd
on those more relevant to evaluation, such as soft and critical
Edition (2016) (with Sjon van ‘t Hof).
systems, viable systems and cultural-historical activity systems.
https://gum.co/wicked
Although I am now based in Aotearoa New Zealand, I have
conducted organisational development and evaluation projects as Systems Concepts in Action: A Practitioner’s Toolkit. (with Richard
well as many systems and evaluation workshops in Aotearoa, Hummelbrunner), Stanford University Press, 2011
Australia, North and South America, Europe, Africa and Asia. https://www.sup.org/books/title/?id=18331
In 2015 I was presented with the American Evaluation Association’s
Video workshop presentations at the Research Institute for Humanity
Lazarsfeld Award for my contributions to the theory and practice of
and Nature. Kyoto. Japan
evaluation – especially the introduction of systems ideas into
https://www.youtube.com/watch?v=lFcWhGE7moQ
evaluation practice.
https://www.youtube.com/watch?v=5RRHpXl2hrw

System diagrams 1
ff
fi
 Why a Book on
In particular:
There is a rich choice of diagram-based approaches. I have noticed
that many people starting out with system diagrams think that there
are only one or two kinds of system diagrams. In fact, there are

System Diagrams?
dozens if not hundreds of di erent system approaches that use
diagrams. Each of them have their strengths and weaknesses. All
have assumptions built into them. There are inevitable debates
about when to use which kind of diagram for what purpose. So I
thought it would be useful to introduce and explain a few that I use
that together span the most common reasons for using diagrams.
Experience matters with using diagrams. The more you do it, the
better you get. There are some tricks of the trade that I can pass on
About system diagrams to you that will help your learning process.
Although there is nothing wrong with working on your own with
Humans have drawn diagrams for millennia. Diagrams have been system diagrams, generally speaking it is better as a collective
used for many purposes, providing visual simpli cations of complex participatory process. Although this is not a book on facilitating
situations and ideas. So perhaps it was inevitable that when the group processes, I have a few tips about how to manage the
modern systems eld began to develop around a century ago participatory side of things that can help make that process as
diagramming was at its heart. And it has continued to be so. constructive as possible.
Constructing diagrams is often an essential part of systems There’s a very substantial prejudice against drawing in many
methodologies and methods. It is also a common way of illustrating cultures. Certainly, I am always a bit anxious when faced with a blank
the ideas that underpin systems methodologies and methods. sheet of paper or an empty screen. I needn’t be, nor should you,
I have been using system diagrams for many years, and have found since there is virtually no such thing as a bad diagram. But it can
their richness useful when designing, conducting, facilitating, help to have some idea of what a good diagram can do.
re ecting on and presenting my work. I have also found them Given the range of possible diagrams to choose from, are some
excellent ways of engaging people to address complex issues. diagram approaches better than others? The answer is yes, but it is
Over the years, I have developed a good understanding of what is not a matter of presenting you with a simple typology outlining which
needed to use system diagrams appropriately and e ectively. So I one is better for which circumstances. I’d like to be able to do that,
thought it would be useful to pass on my knowledge and experience. but in reality the ‘right’ diagram approach is a complex mixture of
purpose, circumstance, timing, resources and personal comfort.

System diagrams 2
fl
fi
ff
fi
ff
 that endeavour.
What’s in the book Cultural-Historical Activity Theory (CHAT). A diagram that
guides a series of questions about how people with di erent
I have selected six approaches to system diagrams. All of them
motivations can work together creatively and e ectively,
originated within the systems eld and have been around for many
especially when responding to changing and challenging
years. Three of them involve gaining insights by constructing
events. It is frequently referenced as Activity Systems.
diagrams from scratch. The resulting diagrams are unique to that
situation and those who developed them. The remaining three
involve gaining insights by applying a generic diagram to a speci c Facilitation
situation.
As you will discover, systemic approaches in general and system
I have chosen these six partly because of my experience using them,
diagram use in particular is best done in groups. Preferably groups
but also because together they span a wide range of potential uses.
that re ect multiple perspectives, viewpoints and ways of
understanding the situation under investigation. Group facilitation is
Constructed Diagrams a book in itself, and it would be easy for facilitation issues to
dominate this one. Consequently, I have only provided facilitation
Rich Picturing. Useful when trying to get to grips with the
advice when it is speci c to the method.
complexity of a situation and work out which aspects you
should be focusing on. Having said that, there is one issue common to facilitation, and
especially relevant to groups using systems ideas. Power. Power is a
In uence Diagrams. Useful when you want to have a snapshot
complex issue, we can consider:
of what’s a ecting a particular issue or topic at a particular
moment in time. power over
Causal Diagrams. Useful when you want to explore the power to
possible consequences of speci c causal relationships within a power with
particular issue or topic over a period of time.
power within

Generic Diagrams These dynamics can become tricky when bringing multiple
perspectives together in a room. Get good advice before doing this,
Cyne n. A diagram that helps you to think about and respond colleagues of mine have had guns put on tables, and I have worked
to a particular situation, or intervention, through ve di erent in situations where people’s livelihoods were potentially at risk.
framings: simple, complicated, complex, chaotic and confused.
In the Introducing System Diagrams chapter, I also discuss another
Viable System Model (VSM). A diagram that helps to assess major issue in facilitating discussion around diagrams, whether the
and ensure the viability of a purposeful activity, given the discussion should be based around consensus or di erence.
range of competing demands from di erent components of

System diagrams 3
fl
fi
fl
ff
fi
fi
fi
ff
ff
fi
ff
ff
ff
fi
 More about this book How to read this book
This is a relatively short book given the ubiquity and sometimes This is intended largely as a workbook not a textbook. This has
complexity of using system diagrams. The descriptions are not various implications:
equally long. Some of them are relatively short and some not so
There are only basic introductions into the conceptual
short. But none is longer than fteen pages. Each chapter contains
underpinnings of the diagram approaches. If you wish to know
what I think is the minimum you need to know in order to use a
more about these well-established systems approaches there
particular approach successfully.
are many excellent references on the Internet.
You will not nd many speci c references in this book, although I
It is far better that you read the book with a particular
generally acknowledge the original sources of my knowledge. If you
application in mind. You can read the book as a text book, but
wish to have more detail, then the Internet has many excellent,
in my experience you will probably miss some important
detailed guides to each of the six approaches I have selected. Given
learnings.
your own particular purposes and the dynamic, changing nature of
the Internet, it is more e ective that you discover them for yourself You don’t have to read the book sequentially from the rst
and use the ones that suit you and your situation. page to the last. In fact, when writing this book I assumed you
may well pick and choose. Consequently, some material is
Books that explore system diagrams in general, or summarise a
repeated in more than one chapter. When that was not
range of methods, or give advice about when to use which diagram
appropriate, I refer to the relevant part of the book. On the
approach are not especially common. At the time of publication
other hand it is a good idea to read the rst three chapters
there is one standout example that has had a substantial in uence
rst. Indeed, the chapter entitled System Diagrams contains
on my thinking and this book. For some years, the UK Open
methods and advice relevant to many of the diagramming
University has made publicly available the resources associated with
methods in this book.
its course on system diagramming (T552). The course material and
videos provide a comprehensive overview of the eld of system
diagramming.
You can currently access speci c resources from these two URLs: Further resources
https://www.open.edu/openlearn/science-maths-technology/
computing-ict/systems-diagramming/content-section-1 A note about references. There is a vast amount of information on
the Internet. Most information refers you to many other sources. It is
http://www.open.ac.uk/stip-resources a very big, constantly changing, rabbit hole. Consequently, I have
focused on a few references that have been around for a while and
are good places to start. All URLs are working as of July 2021.

System diagrams 4
fi
fi
ff
fi
fi
fi
fi
fi
fi
fl
 Why Diagrams?
Why would you want to diagram
An important thing to understand is that diagrams are not only
something or interpret a representations of reality, they help create realities.
• They can summarise complex situations, allowing you to
diagram? appreciate the complexity while seeing the individual components
and the connections between the components.
Consider two possible reasons: • They can give you new insights into a situation by making you
Creating meaning. We engage with a diagramming process or view think carefully and learn more e ectively about the components
a particular diagram to gain a deeper understanding of a particular and connections in that situation.
situation that interests us, a problem that needs resolution or for Any representation of anyone’s reality or imagination can be covered
gaining clarity around a task. Whether working on your own, or with by above description of ‘diagram’. In the English language this would
a group, meaning-making is a process. It is thus critically important include illustrations, patterns, drawings, maps, even pictures and
how that process is constructed and constrained, and who is or is maybe cartoons’. I make no apology about this. Partly because it
not involved in the meaning-making. really doesn’t matter, but mainly because these distinctions are
Communicating meaning. In this case the task is to communicate cultural and linguistic. Some languages do not make the same
aspects of a situation to others so that they can either understand distinctions of meaning – and nor will this book.
your meaning-making or make their own meaning. There are many
jokes about the IKEA assembly instructions, but in many ways they
are remarkably good at conveying visually the necessary tasks to put
together a set of shelves. Here is a nice example from the side of a
tub of co ee machine cleaner:

System diagrams 5
ff
ff
 How about this:
What comes to mind when you
read the word ‘diagram’?
If that is what diagrams do, then what makes a diagram a diagram?
Consider this situation, described in the Open University’s T552
system diagramming course (see reference below):
Jane is married to Tom and they live at No. 8. Tom's sister
Dawn lives at No. 20 and has three children, Peter, Paul and
Mary. Her partner Derek left her 4 years ago and moved to
Scotland, but now she is living with John. John has two
children of his own, Tim and Nicholas, from his marriage to
Julie. Julie's father, Alf, works in the same factory as Tom's
father, while Dawn and Jane's mother went to school together.
Alf and Millie live at No. 34. They used to live at No. 6, but Got it now?
moved after the children left home. Dawn has two older
brothers, one of whom has moved away. Keith and Pamela are Essentially, the diagram replaced a linear description of ‘reality’ that
Tom's parents, but Pamela died last year. Keith now lives alone revealed itself over your journey through the text, by a single image
at No. 18.* that presents the information in an expansive one-shot vista. Short of
a time-warp, oral presentations and writing are linear. Reality is
delivered to you one component at a time. It is like waiting for a
Got it? freight train to go past at a rail crossing – could you describe the
driver in the cab once the whole train had passed by? More
Maybe not. prosaically, have you ever experienced reading an executive
summary and forgotten what was said in the rst paragraph? Have
you had to reread it a couple of times to make sense of it? In a way, a
diagram is an executive summary all in one go.

*This example and others are taken from the UK Open University’s course on system diagramming. Much of this chapter is based on this excellent course and I encourage you to
read the course material. It will only take you a couple of hours, and expands on this introduction. You can access the material here
https://www.open.edu/openlearn/science-maths-technology/computing-ict/systems-diagramming/content-section-1

System diagrams 6
fi
:

And as the situation gets more complicated, the problem of Or this generic diagram of an activity whose components need to be
meaning-making, of making sense of the situation increases. I recall kept in balance in order for it to sustain itself
an interview for a job when I was asked to describe a spiral staircase
while sitting on my hands. Try to describe this complex, dynamic,
performance management situation in words.

Env
3

iron
2 2

men
t
1 1

1 1

1 1

System 1…operative units

System 2…coordination and information
System 3…management and control
System 4…intelligence, monitoring of environment
System 5…policy, coordinates internal and external, present and future

[PRP = Performance Related Pay] Based on Williams & Hummelbrunner. (2010) Systems Concepts in Action. Stanford Press
Source: Open University T552 course

System diagrams 7
 Here is how the Open University T552 course explains this:
Di erent realities, perspectives, Whenever we take in, think and express new ideas, we
framings and purposes describe and represent the ‘reality’ we perceive (in words,
lines, pictures, symbols or numbers) by making simpli cations
for some purpose. In simplifying, we select certain features of a
So diagrams help you ‘get’ the overall picture. But what do you ‘get’? situation – the essentials – to communicate a clear message.
You don’t get ‘reality’ in any of the diagrams. You get a partial The main power in diagrams comes from an individual or
representation of reality. There is nothing inherently wrong with this. group using them to actively further their own thinking.
Diagrams cannot show everything, even the most detailed depict
‘reality’ only partially. But again that is true of any depiction, oral, But let us return to the issue of what you ‘get’ from a diagram.
written or visual. I could probably take a photo of my desk that might That substantially depends on two things:
accurately depict the objects on it, but it would be literally just a
snapshot, a moment captured in time. By the time you read this, a 1. The purpose of the diagram
year or so after I rst typed these words, the desk will be di erent. 2. The meanings that people draw from the diagram
Nor would that photograph explain the meaning or signi cance of a
tiny model Andean llama underneath the monitor. And of course the
llama is itself merely a diagram of the real animal.
Thus …
Diagrams are representations of reality that are designed to serve
a particular purpose and invariably re ect a particular perspective
on reality. Di erent purposes, di erent perspectives, di erent
representations of reality emerge as di erent components of reality
become relevant.
And …
Diagrams are generally about leaving things out to enhance a
particular understanding rather than putting everything in. That’s a
generalisation, not always – as we shall see. And sometimes they
don’t re ect reality either but also help create realities – again as we
shall see. But the general rule is as simple as possible without being
simplistic.

System diagrams 8
ff
fl
ff
fi
ff
ff
fl
fi
ff
fi
ff
 The purpose of the diagram
The T552 course contains three additional diagrams that can be
constructed from the explanation about people’s relationships to
each other.

Each represents a di erent purpose for drawing the diagram.

Show residential history
Show who is related to whom
Show who lives where

Can you identify which is which?

Same data, di erent purposes, di erent diagrams. The issue when
selecting diagram-based approaches is not the approach itself but
whether it ful ls the purpose for which it was selected.

System diagrams 9
fi
ff
ff
ff
 The meaning that people draw
from the diagram
Look at the diagram to the right.

Right now you may be seeing a young n-de-siècle woman. On the

other hand you may be seeing an older woman. You may be seeing
both. If you return to this diagram in a day or so, you may always see
the older woman rst or the younger woman. There is a reason for
this and it is a matter of cognition, of experience and perspectives. A
study revealed that older people were more likely to see the older
woman before the younger woman, and younger people the
opposite. Apparently the ability to identify a face in this diagram is
in uenced by your experience. People recognise faces of their own
age group more immediately than those outside their own age
group. Psychologists call this phenomenon ‘own-age bias’.

The broader message – and a critical one – is that people will re ect
their particular perspectives, framings, history, values and other
aspects when constructing or interpreting diagrams.

System diagrams 10
fl
fi
fi
fl
 To expand on this. Here are two examples of a diagram that seek to
show aspects of a particular intervention to prevent the spread of
malaria in Northern Perú.

Take a close look at the similarities and di erences between them.

Both people who drew their diagram were given the same written
description of the situation and asked to depict that situation visually.
In other words, the data they used were the same. However, from my
perspective, I can see that they chose to emphasise di erent
aspects of that description. In my reading of the diagrams, the top
diagram identi es the long-term and short-term aspects of the
situation in some detail whereas the lower diagram emphasises
current aspects of the situation. The top diagram also emphasises
the impact of the situation on school children whereas the bottom
one doesn’t identify the issue at all. And in both cases the drawers
added things that were not in the written description, presumably
drawing on their own knowledge of similar situations, their values,
assumptions and interests.
It is important to emphasise that neither diagram is right or wrong.
They are just di erent. And my reading of the diagrams will be
di erent from those who drew them or indeed from yours.

System diagrams 11
ff
fi
ff
ff
ff
 So di erent people will look at the same data and make sense of it in
di erent ways. They will draw di erent conclusions. The data do not So what for diagrams?
speak for themselves, words are put in the mouths of data by people
bringing their own perspectives, values, experiences and histories to What do you think are the consequences for using system diagrams
bear. for the issues raised in the last few pages?
You may think that is stating the obvious, but in my experience the Primarily, it implies that diagrams are more than representation of
obvious often ies out of the window when discussing, constructing reality, or even making meaning of reality, it is about making
or observing diagrams. meaning from reality. Indeed, if a diagram does not tell you anything
new or something you did not know beforehand, then maybe that
It is critical in using system diagrams to keep in mind this mantra:
particular diagram has limited use. It is for that reason, when using
Do not assume that anyone else except you will make sense of a system diagrams, you need to consider what is not known about the
diagram in the same way that you do. situation and include that ignorance in the diagram (if it is being
For instance, I circulated an early draft of this book to thirty people constructed) or including in the interpretation if you are observing an
for their comments. Several people disagreed with my choice of existing diagram.
important components in an example I used to illustrate the chapter There is one more important thing to consider:
on Causal Diagrams. They suggested alternative, and in all cases
Be clear what assumptions are being brought to bear. If you are
di erent, components based on their own assessment of the
drawing a diagram then that assessment will contain hidden
situation I had described. As mentioned earlier, diagrams do more
assumptions about the way you see the world, your values, your
than illustrate a situation, they create meaning of a situation to
history, your knowledge, your priorities. If you are interpreting a
people.
diagram developed by someone else then that interpretation will be
All of this challenges the assumption that somehow tidying up a based on your assumptions about the way you see the world, your
diagram will make it more likely that someone else will understand a values, your history, your knowledge, your priorities. These will never
diagram more clearly. Be aware that in this case more clearly may be exactly the same as anyone else’s.
well be a euphemism for the same way as me. As the systems
thinker and practitioner Derek Cabrera observed – you cannot
ignore cognition when working with diagrams.

System diagrams 12
ff
ff
ff
fl
ff
 Why Systems?
Ray Ison’s diagram below illustrates this diversity well:

Hopefully you are a bit clearer now about what comprises a diagram.
Now what does the ‘systems’ part mean?

The systems eld

Let me rst clear up some confusion about what comprises the
systems eld.
Many consider the systems eld to be entirely about methods that
either diagram or simulate behaviour expressed as a series of boxes,
lines and arrows. That confuses the entire systems eld with one
methodology within that eld called system dynamics. Source: Ison R (2010) Systems Practice: How to Act in a Climate Change World. Springer

In fact, the systems eld contains a very large range of On the left hand side of the diagram there are the intellectual
methodologies, methods and techniques. For instance, not many traditions that have contributed to the development of the systems
people know that the ideas underpinning family therapy emerged eld. Next some names of people who helped that development. In
from the systems eld. In the social sciences, anthropologist the middle is Ison’s attempt to group them into what might be
Margaret Mead is widely acknowledged as a key gure in the regarded as system disciplines. Further to the right are some of the
systems eld. A large chunk of management science and information speci c methodologies and methods that have emerged from those
science are based on early systems concepts. In fact, if you hear disciplines.
someone talking about ‘systems theory’ as if it were a single Of course this is a diagram of a system of systems ideas!
construct, then encourage them to look at this website: Consistent with my earlier remarks about diagrams, Ison
https://en.wikipedia.org/wiki/List_of_types_of_systems_theory acknowledges that this is just his conceptualisation and that there
are others.

System diagrams 13
fi
fi
fi
fi
fi
fi
fi
fi
fi
fi
fi
fi
What is a system? 1. A system is an assembly of components connected
together in an organised way.
First, let me explore the vexed issue of what is a ‘system’. The stress here is on the word ‘organised’. This means, for example,
The Open University has run courses on systems, systems thinking that the components of a clock piled together in a heap on a table
and systemic practice for over fty years. In that time it has awaiting assembly do not constitute a system. After assembly,
developed a sophisticated de nition and explanation of what however, they do make up a system because their interconnections
comprises a system, which I have reproduced below from the T552 have been organised.
course material:
=====
1. A system is an assembly of components connected together 2. The components are a ected by being in the
in an organised way. system and are changed if they leave it.
2. The components are a ected by being in the system and are
As an example, think of the physiological system that is your body. Its
changed if they leave it.
components are organs such as your heart, liver, kidneys and lungs.
3. The assembly of components does something. Clearly they are changed if they are removed from the system: they
decay, they become inert and lifeless.
4. The assembly has been identi ed by someone as being of
interest. So far, so good. But, actually, this part of the de nition has very far-
reaching implications. For if the components are di erent in some
This seems a straightforward list, but each component has important
way when they're in the system from when they're not, then the
implications.
system as a whole must have properties deriving from the
Let's examine them in turn. interactions of the components that cannot be deduced or predicted
from an inspection of its components taken in isolation. One
common way of putting this is to say ‘the whole, the system, is
greater than the sum of its parts.*

* I prefer to use Derek Cabrera’s concept that a system is the sum of its parts and the relationship between the parts (Pers. Com.)

System diagrams 14
ff
fi
fi
fi
ff
fi
ff
 3. The assembly of components does something. a boundary that distinguishes those bits of what you ‘observe’ from
everything else – the system's environment (see below). Hence you
This part of the de nition covers two aspects. First, systems are cannot think of a system without thinking of its boundary and what is
dynamic entities. Second, and even more important, that a system outside the boundary (its environment).
behaves as if it has some purpose.
The concept of environment might seem reassuringly familiar to you:
Thus … it is not enough to talk about the telephone system as if it it is a term in daily use – such as when people talk about ‘pollution of
were just a thing that everybody instantly recognises and fully the environment’. But the everyday interpretation of the term, which
understands in the same way as, say, a pencil. Rather, we need to usually amounts to ‘everything out there’, is too fuzzy, too imprecise
talk about a ‘system for employing electrical engineers’ or a ‘system to be useful. So we adopt a more restricted interpretation:
for enabling friends and family to chat to each other’ or ‘a system to
The environment of a system is made up of those things that are not
give investment returns to shareholders of the telephone company’.
part of the system, but can a ect the system.
Notice that these descriptions uses active verbs to highlight its
dynamism and purpose and so make it more speci c and The concept of a system boundary follows naturally from that of
identi able. They are not descriptions of something, but description environment. It is simply a notional line that divides a system from its
of what something does. environment. This means that a boundary is a subjective idea.

4. The assembly has been identi ed by someone as

being of interest to someone.
This is important because it brings out the fact that a system is
essentially a private, personal idea: a construct. A system can be
(and often is) simply a personal ordering of reality, the result of
seeing some degree of orderly interconnectedness in some part of
the world. Thus a system, any system, can be many di erent kinds of
system simultaneously, depending on who is studying it and why, as
shown immediately above. In some cases we might agree on the
purpose of the system, in other cases we may have to deduce a
purpose for the chosen system.
This de nition of a system is incomplete, however, without
considering two other, related, terms: environment and boundary.
Source: Open University T552 course
The reason for this is that you can only de ne a system by ‘drawing’

System diagrams 15
fi
fi
fi
ff
fi
fi
fi
ff
 might I describe a systems approach? Generally, I use it to describe it
as an entwining of two activities, systems thinking (a cognitive
A system boundary doesn't have to correspond to any real-life
activity) and systemic practice (a physical activity)
barrier or the limits of an organisation, for instance.
Not surprisingly, given the focus of this workbook, I am going to
Choosing where to draw a system's boundary therefore requires you
explain this using a diagram. And consistent with the discussion on
to exercise some judgement. It also depends very much on your
systems, you should view it as my description of systems thinking
reasons for examining the system. Thus, if you're interested only in
and systemic practice and not the description of systems thinking
the reason why a particular component in your car has failed, it may
and systemic practice.
be appropriate simply to draw the boundary round the engine or a
particular sub-System of parts (e.g., the electrical system). But if you The combination of systems thinking and systemic practice can be
are a government safety o cial investigating why a particular understood as having three elements.
component tends to fail in a certain make of car, you may have to
widen the boundary to include the dealer-network for the marque in
question, the factories that manufacture the model, suppliers, Understands interrelationships
subcontractors, trade unions, etc. Knowing where to draw the Engages with perspectives
boundary is one of the ne arts of systems practice. Where you
choose to draw it can often say a great deal, not just about your Re ects on boundaries
views, but also about your values.
=====

So what is a systemic approach?

I am very comfortable with the Open University’s description of a
system. It tells me that a system is not only a material ‘thing’, but also
a human construct. You and I may look at exactly the same reality,
but because we di er in our histories, values, motivations and
interests, you will probably draw di erent boundaries to me and
emphasise di erent systems components. Crucially, you may even
see the system serving di erent purposes than me. I call this
di erent ways of framing a system, a word that will occur frequently.
So within that understanding of what constitutes a system, how Let me expand on the diagram.

System diagrams 16
ff
fl
ff
ff
fi
ff
ffi
ff
 Firstly, any endeavour has to look at what is really out there relevant But of course you cannot collect every single piece of reality that is
to the endeavour (understands interrelationships) and gain an available, from atomic particle interaction to global dynamics, or
understanding of how di erent people interpret that reality (engages include every person on the planet. The idea that holism is about
with multiple perspectives). including everything is a theoretical and practical impossibility; it is a
much misunderstood and frequently misused term. Every systemic
However, there is so much reality that might be relevant and there
endeavour will exclude something or somebody, every systems view
will be hundreds if not thousands of di erent perspectives on that
is partial. Rather than being about everything, holism is about being
reality. So if we actually want to think about things, or indeed actually
very smart, informed and thoughtful about where to draw the
do something as a result, there is a real risk of being frozen into
boundaries around your task. Thus the choice of who or what to
immobility, like a deer in the headlights. To be able to think about
include and who or what to exclude in a system diagramming
something, or do something, you have to draw boundaries. You have
process is a boundary decision that must be able to claim that
to decide what reality and perspectives to leave in your systemic
whoever ends up being included are the right people, with the right
inquiry and what to leave out so that you can arrive at thoughts and
knowledge and the right perspectives for the task. And who or what
actions that are both feasible and desirable. And having drawn them
is excluded have their interests fairly considered.
you need to re ect on the implications of those boundary decisions.
This boundary setting is not just about physical things like space or
objects or people. It also includes decisions about intangible things
such as purpose, time, knowledge, agency, and especially values
and beliefs. They can be a zone rather as well as a de nitive sharp
frontier. They can be leaky – allowing for things adjacent to the
boundary to move between the system and environment.
Ultimately however, you have to be con dent that you have arrived
at a point in your thinking that you can be su ciently con dent that
you can at least make a viable claim that this is a right way to think
about things or a right thing to do. You have to reassure yourself that
you have taken the most appropriate aspects of reality and the most
important perspectives into consideration. And whilst doing so you
have attempted to gain su cient legitimacy to be able to claim or
defend that this is the right way to think about or act within the
situation. This involves a lot of talking and listening. For that reason
many systemic practices, including system diagramming, tend to be So now let’s consider this in terms of system diagrams in general
participative. It is an important part of seeking to explore su cient terms.
amounts of reality and an adequate range of perspectives to be
con dent in the boundary choices.

System diagrams 17
fi
fl
ff
ffi
ff
fi
ffi
fi
fi
ffi
Introducing
Is it a depiction of a system?
No, it is not necessarily a diagram of a system. Indeed in some cases

System Diagrams:
regarding the diagram as a system is counter-productive.
That’s because most system diagrams are not stand alone ends in
themselves. They are means to a systemic end. Most have been
developed as components of particular systems methods and

Why, Wha ,
methodologies, or they seek to depict a systemic method or
methodology. They are devices to be used as part of a systemic
inquiry, they are not the end product of a systemic inquiry. Often they
are merely the start. In those terms whether the diagram itself

When, Who and

represents or can be described as a system is neither here not there.
Rather the diagram is a stage in the application of systemic ideas in
order to gain deeper insights and possible strategies for further
inquiry or problem resolution.

How These distinctions are important because if you consider all system
diagrams only as representing a system, then for some of those
diagram approaches you risk trapping yourself in the very set of rigid
constraints and understandings of a situation that you are trying to
avoid.
For instance, the Rich Picture illustrated on the next page is not a
diagram of a system. Indeed, the systems methodology from which it
originated speci cally states that it should not represent a system.
Having explained what a systemic approach comprises and the One the other hand, the Causal Diagram on the page after would
nature of diagrams, what exactly is a system diagram? comply with the de nition of a system. Yet both are useful in
systemic thinking and practice.

System diagrams 18
fi
fi
t
 At one extreme is Rich Picturing …
Two distinctly different kinds of
system diagrams
Constructed Diagrams
Constructed Diagrams are assembled from scratch, based on
speci c purposes, established procedures, speci c rules and the
kind of data available. Each diagram is unique to those who
constructed it and to the situation it re ects.
In this book, three kinds of Constructed Diagrams are explained.
Rich Picturing
In uence Diagrams
Causal Diagrams
To some extent these three chapters are ordered in terms of
increasing focus, speci city of purpose and understanding of the
situation.

… where in many cases you are exploring the unknown. It can be

unknown to all those participating in the construction, or it can be
unknown to some participants according to the perspectives they
bring.

System diagrams 19
fl
fi
fi
fl
fi
 On the other extreme, are causal diagrams…

Generic Diagrams
Generic Diagrams are in many ways the polar opposite of
Constructed Diagrams. In a Constructed Diagram you assemble a
diagram that emerges from a speci c situation using general rules
and processes. In contrast, Generic Diagrams are existing diagrams
that you apply to a speci c situation using speci c rules and
processes.
The three Generic Diagrams in this book are:
Cyne n
Viable System Model (VSM)
Cultural-HistoricalActivity Theory (CHAT)
These three diagram types are visual representations of a speci c
… which often need to re ect a very speci c purpose, a single systems approach. They di er from other systems approaches in
framing and a good deal of knowledge about the dynamic that they depend very heavily on their diagram to describe how the
relationships between the components of the diagram. ideas within that approach can be applied. All three of these
approaches are very sophisticated. This book can only cover the
In uence Diagrams sit somewhere in between. basics of how to apply these approaches, but I encourage you to
It is possible to start with a Rich Picture and move to Causal explore the many in-depth resources on the Internet. Having said
Diagrams through In uence Diagrams, although I’d recommend that, in each chapter I list a series of questions that ow from the
some caution since they are based on di erent systems approach. You should be able to use those questions without a
methodologies. If you are sure that there is only one relevant detailed understanding of the particular approach.
purpose and already have a clear idea of the dynamics, then you As with Constructed Diagrams, the key reason for Generic Diagram
could go straight to Causal Diagrams. selection is the purpose of the systemic inquiry. However, in
Constructed Diagrams the choice depends largely on the purpose
for which you are diagramming and the degree of knowledge of the
situation. In Generic Diagrams the purpose is more hard-wired into
the approach itself. In the case of Cyne n – responding to
complexity, VSM – sustaining viable systems, and CHAT – learning.

System diagrams 20
fl
fi
fl
fi
fl
ff
fi
fi
ff
fi
fi
fl
fi
 Selecting the boundaries of
system diagrams Involvement Stakes Space
Scale People Knowledge
There are three issues to be addressed: Purpose Time Control
Values Beliefs
What are the important boundary choices when using system
diagrams?
Whose interests ought or should to be included in the
construction or use of a system diagram?
Whose interests ought or should to be included?
Who ought be involved in the construction or use of a system
diagram? It should now be clear that not every interest or perspective can be
accommodated or involved. Some important boundary decisions
It is worth noting that this book frequently uses the word ought. In have to be made.
the systems eld, ought is commonly used to imply a necessity or an
imperative. It uses should when implying an opinion or an option. Critical Systems distinguishes between those interests (people,
ideologies, values, resources) involved in a system and those
interests that are not involved but are a ected by a system. It argues
What are the important boundary choices when using
that when drawing boundaries about issues, people and human
system diagrams? constructs, like sustainability knowledge and legitimacy, that not only
should aspects directly involved in a system be included but also
There is an area of the systems eld, Critical Systems, speci cally
those a ected by a system but not directly involved in it. In other
concerned with boundary decisions.* As I described in the previous
words, those aspects that, by being outside the system, are
chapter, a boundary choice is necessary in everything we do. We
e ectively marginalised or victimised by the system. C. West
cannot do or think about everything. To actually do anything we have
Churchman, who developed Critical Systems, provocatively, although
to exclude something. Usually that’s lots of things. Critical Systems
not entirely inappropriately, called them the enemies of the system.
requires us to deliberate on who or what bene ts and who or what is
Indeed, he wrote a book The Systems Approach and its Enemies
harmed by those choices so that we can make claims that whatever
that makes this stance explicit.
it is we are doing is the right thing to do. And so it is with diagrams,
whether developing them or interpreting them. We have to
deliberate on and decide what to include or exclude in terms of our
choice of at least the boundaries opposite:

* https://wulrich.com/downloads/ulrich_2003b.pdf

System diagrams 21
ff
ff
fi
fi
ff
fi
fi
 Marginalised, victimised, enemies. These are strong emotive words.
The reason Churchman argued that their interests be acknowledged How to decide key boundaries
by the system so that the system designers can claim that what the
system does or what they wish it to do is indeed (or can defended
as) the ‘right thing to do’. But how do you do that, given that, as How do you do this?
pointed out in an earlier chapter, you cannot include the universe
and its multitude of interests, known and unknown?
Stakeholder analysis
Werner Ulrich addresses this problem in two ways in his Critical
Systems Heuristics*. Firstly, focus only on those aspects that are not Most frequently I do it through a form of stakeholder analysis. Below
only marginalised but also negatively a ected by the system. is an example. The analysis is never quite the same every time, but in
Secondly, to ensure that in any deliberation of the systems general it can be illustrated by this example of rice production in the
boundaries these interests be represented and acknowledged in Sahel region just south of the Sahara desert (see also the Rich
some way. This inclusion could range from as little as being su cient Pictures chapter). For clarity I have only lled in one item per box, but
to argue that marginalising these interests is appropriate to as much in reality there will be multiple entries per box.
as sweeping-in aspects of those interests to su cient legitimacy of
the system to do what it wishes to do. Think of it as a kind of
systemic Greek Chorus. Of course, if you chose to sweep-in some of Involved in
those marginalised issues, you probably now have a new set of
marginalised issues to contend with. Interest,
The role is The role
Stakeholder role motivation, or
a ected by ... contributes to ...
stake
Who ought to be involved?
Acquisition of
Consequently, in terms of who should be involved in the construction Village leader Election process Authority
resources
or use of a system diagram, then the above Critical System principles
should be the guiding force. Obviously, if the diagram is being done Maintenance of
Farmer Village wealth Village economy
by a small unrepresentative group then that group has to work out irrigation
how these various excluded interests can, or ought to, be swept into
Fertiliser
the work. Fertiliser seller Rice productivity Pro tability
availability
However, if the diagram construction or use is participatory, then the
Village Future Village Future
decision of who can most e ectively re ect the various interests is
critical. Who ought to be involved in that process?

* See https://wulrich.com/csh.html

System diagrams 22
ff
fi
ff
ff
fl
fi
ffi
ffi
 Negatively what motivates them and so on. In most cases there will be more
a ected by than one item in this box.

Interest, You may have realised that this kind of framework is useful not only
The role is The role to help select what interests need to be included in the diagram
Stakeholder role motivation, or
a ected by ... contributes to ... approach, but also the framework can be used during the entire
stake
investigation. That includes what is and is not included in a diagram.
Cost of locally
Demand for In particular, as a result of developing or using the diagram, new
grown rice
Rice consumer locally grown Cost important stakeholder roles may emerge, or new dynamics of what
relative to
rice a ects them and what they a ect. In which case, you may want to
imported rice
move the non-human stakes to the rst column and begin to identify
Health of what a ects those stakes and what those stakes contribute to.
Environmental
waterways
Health Obviously, with a requirement to include some way for interests of
downstream.
those negatively a ected by the current situation or proposed action
Once I have assessed the interest, motivation or stake (or whatever needs careful thought. And also the risk involved in bringing
boundary I decide to include in the analysis) then I can deliberate on together di erent interests that may have quite di erent stakes in
the consequence of including or excluding them in a diagram activity. the situation. For instance, in a country where it was fairly risky to say
things that ran counter to the ideas of the dominant elite, we held a
Notice that I identify two kinds of stakeholders. One is identi able
series of separate workshops rst, and then a summary workshop
people (or groups of people) and the other is a non-human feature of
where those di erent contributions could be explored in a smaller
the situation. Also note that I also use the term ‘role’. That’s because
group in a safer way.
a single stakeholder (even non-human ones) can often occupy
several roles in a particular situation, such as a village leader and a
farmer. It is important to understand that how they juggle the
possibly reinforcing and possibly contradictory roles will a ect how a
situation behaves. In practice, especially when doing this in a group
setting, the non-human stakeholder or even absent human ones
need to be represented in some way. In this case example the non-
human stake was represented by a rural sociologist.
Generally speaking, the table is completed by working from the left
to the right. Once the stakeholder roles are identi ed, then what
a ects that role and what it, in turn, a ects or contributes to can be
identi ed. From exploring these two you can identify what is at stake
for that particular stakeholder role, what skin they have in the game,

System diagrams 23
ff
ff
ff
ff
fi
ff
ff
ff
ff
ff
fi
fi
ff
fi
ff
ff
fi
 Acknowledging multiple perspectives
What questions should diagrams What issues are being expressed by key actors, what interests,
pose? stakes, or motivations does that reveal?
What con icts, tensions, or agreements emerge from these
multiple perspectives?

In each of the diagram chapters, I list the kinds of questions that ow What assumptions underpin or are expressed by the diagram?
from the speci c diagram. However, in addition to these diagram What are the di erent ways a situation can be described,
speci c questions, some questions are relevant to any system viewed or framed?
diagram. These are inherent in any particular systemic inquiry that
uses diagrams. How do boundary issues a ect perspectives?

Here are some suggested questions. It is important to understand How do interrelationship issues a ect perspectives?
that they are heuristics. In other words, they need to be modi ed to
match your particular circumstances. There are two sets: one applied
Re ecting on boundaries
to the diagram itself, and the other about the process of working with
the diagram. What normative issues (viewpoints, opinions, values) are
displayed?
Applied to the diagram Who or what is advantaged or disadvantaged in the situation?
Understanding interrelationships Where is the locus of control in the system? Who or what has
the ability to radically change the situation?
How do the various elements of the diagram interact?
Are certain kinds of knowledge accommodated or ignored,
What are the major processes by which they interact?
and what are the consequences?
What are the interrelationships between resources identi ed in
Who has power or authority to say that what is happening in
the diagram (i.e., people, money, things, knowledge, skills)?
the situation is ‘good’ or ‘bad’?
Which relationships in the diagram are known and certain and
How do interrelationship issues a ect the re ection on
which elements are unknown or uncertain?
boundaries?
How do boundary issues a ect interrelationships?
How do perspectival issues a ect interrelationships?

System diagrams 24
fl
fi
fl
fi
ff
ff
ff
ff
ff
ff
fl
fi
fi
fl
 Applied to the process How should diagrams be
What sense do you make about the diagram or diagrams. What
assumptions are involved in that sense making? interpreted?
What stands out as new insights about the diagram and its use
that those drawing the picture hadn’t seen before? Each of the chapters describe a process of interpretation speci c to
that diagram type. However, there are two important general
What additional aspects were added or modi ed or removed principles.
during group discussions? For what reason?
Recognise that everyone’s interpretation will be di erent.
If the process involved drawing or using more than one They may even frame things completely di erently.
diagram, what is the same and di erent between the di erent
diagrams? Seek dialectic as well as consensus.

Which were signi cant areas of agreement and disagreement If you have read the earlier chapters, then the rst principle should
between those who were involved in drawing or using the be self-evident by now. The second is worth repeating.
diagram? How were those di erences addressed? When interpreting diagrams, especially diagrams that are
How was the process a ected by your boundary decisions? constructed by groups of people, there is often a strong urge to nd
consensus around the diagram and its meaning. However, many
systems approaches speci cally encourage and indeed are based
And … on generating and using dialectic (e.g., CHAT– see the nal chapter).
Essentially, dialectic is the process of bringing together and learning
from di erence rather than similarity. This is why, generally speaking,
If someone from Mars who knew nothing about the situation looked if I am facilitating a group process of diagram construction or
at the diagram or listened to your conversations, what do you think interpretation, I use more than one diagram and more than one
they would highlight as issues? framing. If I am constructing diagrams I will generally seek to
Answer this question starting with “Something to do with …”. This is a produce three or four diagrams. I then get people to look at the
framing exercise that involves look at the diagram thorough di erent di erences between the diagrams and ask them to explain those
lenses, di erent ways to understand a particular situation. Generally di erences but not explain them away (as often happens in a
speaking I encourage people to identify at least three di erent ones. consensus approach where di erences are often marginalised).
When I am using a generic diagram, I get groups to interpret the
diagram separately and then explore the di erences and their
meaning.

System diagrams 25
ff
ff
ff
ff
fi
ff
fi
ff
ff
ff
ff
ff
fi
fi
ff
fi
ff
ff
ff
fi
fi
 expressions of strengths, weaknesses, opportunities and threats.
Now what? What next? There are many variations that can be viewed on the Internet, but
basically a Force Field Analysis poses four questions:
Diagrams are almost never an end in themselves. The results of a
What forces help create a desired state?
diagram-based analysis are generally applied to the resolution of
some identi ed issue or problem. If a diagram is already part of an What needs to be done to help those helping forces help?
existing systems approach, as is the case with four of six diagrams What forces are hindering the creation of a desired state?
types in this book, then it gets plugged into the next stage of that
diagram approach. But it is often possible to use the diagrams What needs to be done to help those hindering forces hinder?
independently of the associated approach. In which case what are
some good ways of arriving at the insights into the now what and Keep, Chuck, Change, Create
what next questions?
This is another simple framework that was originally developed in a
A quick search of the Internet for organisational change processes slightly di erent form for Search Conferences, a large group problem
will yield many useful ways of indicating now what and so what. solving process developed by Fred and Merrilyn Emery during the
Below are three very simple approaches that have long histories in 1950s.
the systems eld. I use them on almost every job I do.
Once you have identi ed the problematic situation or a desired
answer you ask four questions.
What needs to be kept?
How to identify ‘Now What’ and What needs to be chucked?
‘What Next’ What needs to be changed?
What needs to be created?
Force-Field Analysis These four questions may seem totally self evident, but my own
Force Field Analysis was developed by Kurt Lewin in the 1940s as a experience in helping people solve problems is that the rst is often
way of assessing what forces are helping a desired change and what ignored. That not only risks throwing the baby out with the bathwater
forces are hindering a desired change. It can, of course, be used to but leads to poor planning and the failure of many change
explore the opposite, what is helping and sustaining an undesirable processes. In any change process there are substantial aspects of
state and what forces are working against that undesirable state. situation that need to be continued. Identifying what needs to be
Force Field Analysis is often confused with SWOT analyses. In my chucked prevents overload. ‘Changed’ acknowledges that things
experience, Force Field Analysis is superior in many ways to SWOT can be built on existing foundations.
because it explores dynamic forces rather than the rather static

System diagrams 26
ff
fi
fi
fi
fi
The 3 Rs
The 3 Rs questions come from a variety of di erent systems and
organisation change approaches. Some use more Rs and some use
other letters, such as McKinsey’s 7S framework. These questions can
be used in conjunction with the two previous methods.
Who plays important roles in this situation? How would they
have to be altered to address the identi ed issue and achieve
the desired result?

What are the implicit and explicit rules in this situation? How
would they have to be altered to address the identi ed issue
and achieve the desired result?

What kind of relationships are important in this situation? How

would they have to be altered to address the identi ed issue
and achieve the desired result?

System diagrams 27
fi
ff
fi
fi
 CONSTRUCTED
DIAG AMS
Resources
Rich Pictures
There is a lot of material on the Internet about the next three
diagramming approaches. I encourage you to explore and
experiment.
My key resource for these chapters is the Open University’s System
In uence Diagrams Diagramming course T552. The written and video presentations are
excellent. The full reference is:
Open University (2000) Systems Thinking and Practice:
Diagramming (module code T552) OpenLearn platform.
The links below all worked as of July 2021, but the platform is
undergoing change and so may some of the URLs.
Causal Diagrams
https://www.open.edu/openlearn/science-maths-technology/
computing-ict/systems-diagramming/content-section-1
http://www.open.ac.uk/stip-resources

System diagrams 28
fl
R
 Rich Pictures
Purposes of Rich Pictures
In a nutshell, Rich Picturing is about exploring and interpreting the
complexity of a situation of interest in a free wheeling way. It is NOT
intended to be a diagram of a system, or indeed guide you towards
the resolution of speci c issues. Rich Picturing is a tool that provides
a rst stage in helping you decide where and how you might nd
some clues about how to address problematic situations. It can help
you frame the situation in ways that will help you, if you wish, to
develop a diagram of a system. Indeed it was developed as part of
the rst two stages of Soft Systems Methodology*, where the system
diagram is developed several stages later. In contrast to a system
diagram, a Rich Picture is intended to be as unstructured as possible.
There is a common idea about the negative consequences of
rushing to solutions. Rich Picturing steps even further back and tries
to prevent you from rushing to problems. Problems are human Consequently, Rich Picturing in general is best undertaken at the
constructs based on perspectives. What might be a problem to you beginning of a systemic inquiry. In many cases it is useful as a means
may not be a problem for me. Rich Picturing basically says, let’s of identifying the focus, framing or primary perspective of a systemic
forget about problems and solutions for a while and just describe inquiry.
what is there from multiple perspectives. After that we can work out To help people understand this better, I sometimes describe Rich
what problems and solutions may be present within the current Picturing as an alternative to a 15,000 word situation report.
situation from those di erent perspectives.

* https://en.wikipedia.org/wiki/Soft_systems_methodology

System diagrams 29
fi
fi
fi
ff
fi
 Below is a Rich Picture of a situation of rice production in the Sahel
What does a Rich Picture contain region of Mali that depended for its success on some ageing pumps
supplied originally by an Australian aid agency which no longer had
and what does it look like? an interest in this project. The puzzle was what to do when the
pumps nally failed to work. Inevitably, as the Rich Picture showed,
A Rich Picture intends to represent visually as many as possible of the puzzle was a complex one to address
the following aspects of the situation of interest:
the structure of the situation;
the processes between elements of that structure;
important aspects of the situation that a ect how the
stakeholders, stakes, structures and processes interact;
the nature of the interrelationships (e.g., strong, weak, fast,
slow, con icted, collaborative, direct, indirect);
purposes, aspirations, and goals;
motivations;
values and norms;
environmental aspects, e.g., a climate of opinion;
issues, con icts, and agreements;
resources (e.g., people, money, tools, skills);
things you don’t know or puzzle you.

Critically, Rich Pictures should also demonstrate how and where

things don’t interact and whether that is a good or bad thing.

System diagrams 30
fi
fl
fl
ff
Once again, the important thing to understand is that you are not
drawing a system, you are doing the opposite. Systems are
structured. In Rich Picturing you are drawing the unstructured
messiness of the situation that will allow you to think systemically
about that situation.
Thus, Rich Pictures are a means for moving from a state of messy
confusion, where all you know is that you’re dealing with a
problematic situation, to a state where you’ve identi ed one or more
themes that as a group or individual you want to address.
Rich Pictures can cope with any sort of chaos. It happily receives
whatever chaotic mess of thoughts and perceptions pours down
your arm from your brain, out of the pen in your hand and onto
whatever surface you are using to draw the Rich Pictures.
Generally they are hand-drawn and do not require artistic talent.
People are often drawn as stick gures (they certainly are in mine),
relationships as arrows. Labels and text balloons are added to clarify
meanings of lines, arrows and illustrations.
In the end, a Rich Picture can be quite messy and confusing. This
should not put anybody o from drawing them. Often, the messiest Source: Water in the Waimea Basin: Community Values and Water Management Options.
Winstanley, Baker, Foote, Gregory, Hepi, Midgley ESR, New Zealand (2005)
pictures are the richest and most meaningful to participants.
Opposite is a wonderfully messy example of a water catchment
management situation in Aotearoa New Zealand. Whatever it looks like, it is very important that the Rich Picture
conveys all the important elements of a situation without overly
imposing your own understandings and prejudices.
Indeed, Rich Pictures are often drawn before you know clearly which
parts of a situation you should be focusing on. Therefore, when
drawing a Rich Picture, free your mind as much as possible from any
preconceived ideas you may have about the situation. People
drawing Rich Pictures for the rst time often try to place too much
order too quickly into a situation.

System diagrams 31
ff
fi
fi
fi
 How to draw a Rich Picture

Preparation
Ideally, the rst thing to do is to bring together a bunch of people
with di erent perspectives on the situation. You may also do it on
your own – I often do as a way of mustering my own thoughts. It just
risks being less rich.
Once you’ve gathered whoever is going to work on the Rich Picture,
all you will need, other than refreshments, is a large sheet of paper
some felt tip pens, and perhaps pads of Post-its. Don’t have the
piece of paper too big. Two sheets of A3 sized ip chart paper taped Source: Working for Water Programme in South Africa (adapted from Martin Reynolds, Open
University, UK)
together is usually enough. The most common semi-portable white
boards are often a perfect size and an excellent medium for drawing
Rich Pictures, since ideas can be added, deleted, moved around with
great ease. It also conveys the inherent uidity of Rich Pictures. But
Step One – Title
don’t forget to take photos.
If it is a group situation, it is a good idea to have someone to agree The big question that worries everyone when starting to draw a Rich
to take notes of the discussion during the drawing process. As in Picture is where to start. In my experience the best way is to give
most diagramming, the process of constructing the diagram is as your situation a title (e.g., micro-beer distribution in Uganda), then
important as the product. The notes provide a useful memory jogger write that title at the top of whatever surface you are using.
when later analysing the Rich Picture. However, there is a trick to writing that title. Do not phrase it as a
problem, a goal or a question (e.g., how can I get beer from Kampala
to Entebbe?). Rich Pictures are actually trying to stop you rushing too
quickly to a problem statement (implied by the question) or solution.
Ideally it should be a state of a airs without any implied problem or
solution.

System diagrams 32
ff
fi
ff
fl
fl
 Step Two – Start Step Three – Draw

Then the best thing you can do is just start. It really doesn’t matter If you are facilitating a group process, then once people are under
where. However, sometimes those involved need a bit more way your main task is to keep people drawing while they are talking.
structure. Here are some ways of providing that initial structure. It is not uncommon for people to just talk and not get around to
drawing, especially early and later in the process.
Use a related warm-up exercise to allow people to get to know each
other and feel comfortable in throwing ideas around. People will possibly feel uncomfortable and uncertain during the
process. It can be hard to face other people’s often surprisingly
My colleague Judy Oakden likes to start o with a guided or
di erent assumptions because this make us question our own
facilitated conversation about the topic. This can help the group
assumptions. It can mean throwing away the solutions we thought
move more quickly into the areas they don’t understand or puzzle
we had, going back to the beginning and starting afresh. This can be
them. It also helps establish the group dynamic. Deciding the title is
demanding and unsettling (and why a warm up exercise can be a
one possible way of promoting this discussion.
good idea), but that’s often exactly what is needed at the start of a
Another way is to ask people to think individually about the situation systemic inquiry. So if you are facilitating the process, reassure
or issue, write on separate Post-its (in all, maybe 20–50). You can participants that these feelings are normal during the process, but it
use the stakeholder framework in the System Diagram chapter: will be worth their while.
Another common thing that happens is that people start drawing the
Who or what are the key stakeholder roles? picture on a piece of scrap paper (often individually), believing that
the big drawing is for the ‘perfect’ version. The easiest way of
What are the key stakes (e.g., purposes, motivations, values, avoiding this problem is to ensure that there is nothing is within
norms, aspirations, goals)? reaching distance except the items needed to draw the Rich Picture.
In terms of actual drawing, I nd it useful to encourage people to
start with only one drawing implement between them. Otherwise you
Then as a group place the results on the surface you are working on
can nd that two or three unconnected diagrams are being drawn.
in a way that allows them to display the structure of the
You can relax this advice once things are underway.
interrelationships of stakes and stakeholders within the situation.
Once people are happy with the overall con guration, they can I also nd it useful to remind people that lines may represent a
remove the Post-its one by one and start drawing. relationship, but they do not describe a relationship. Every line
needs an explanation or word to identify the nature of the
But often the best thing is just to start drawing.
relationship.

System diagrams 33
ff
fi
fi
fi
ff
fi
 While I encourage arrows to indicate the direction of any
relationship, I discourage lines with arrows both ways. Relationships Step Four – Re ect in action
are not equal, so draw two lines and explain each one. Three of the
diagrams in this chapter have double headed arrows. Look at them A Rich Picturing process will last around 60 to 90 minutes in a group
and ask yourself if the relationships are likely to be equal. session.
If facilitating I nd it useful to push the pause button twice.
The rst time I remind people of what the Rich Picture seeks to
cover. I nd it useful to have the list of the ten key thematic areas a
Rich Picture needs to illustrate. And in particular highlight the
important aspects in blue.
Structures
Processes
Climate, viewpoints, values, norms
Resources (i.e., people, money, things, knowledge, skills)
Issues that people express
Assumptions that are being made
Con ict, tensions, agreements
Purposes, aspirations, goals, motivations
Geography
Unknowns or puzzles

The second pause is to encourage people to focus for a while on

Source: Open University T552 things that they don’t know or that puzzle them. After all, one of the
main bene ts of Rich Picturing is gaining new insights

System diagrams 34
fl
fi
fi
fi
fi
 Product Questions
Step Five – Re ect on Action
How well have the Rich Pictures re ected the situation in terms of
the ten thematic areas (see previous page)?
If this is being done as part of a large group process where several
Rich Pictures are being produced, it is a good idea to start the Which relationships between the elements in the Rich Pictures are
re ection process with a promenade, where people walk between known and certain and which elements are unknown or uncertain?
di erent diagrams. It is also a good idea to have one person What is in the Rich Pictures that you did not expect? What is not in
remaining with each diagram to explain the diagram to people. the Rich Picture that you thought might have been? What might this
People can then return to make any new additions or changes to mean about your assumptions about the situation?
their own diagram. As part of this you may wish to ask each group to
identify three issues that stand out from their and others’ diagrams If someone from Mars, who knew nothing about the situation, looked
at the diagram or listened to your conversations while drawing the
Then consider these questions. As you go through these questions, Rich Picture, what do you think they would highlight as issues?
think what your answers imply for understanding and addressing the Answer this question starting with “Something to do with …”
issues that motivated the Rich Picture.
For each ‘something to do with .’
These questions assume that there is more than one Rich Picture
produced in a group session. Who bene ts and who loses when viewing the situation shown
in the Rich Picture solely through that lens or framing?
Who controls what in this situation? What knowledge is used,
Process Questions by whom, to do what?
Was drawing the Rich Picture easy or di cult? What is it about Where does the power to do things lie? Where does the power
the situation that made it easy or hard? What is it about the to allow things to happen lie?
perspectives of those involved that made it di cult or hard?
If more than one Rich Picture was produced, what are the
Which were signi cant areas of agreement and disagreement similarities and di erences between them? What does that say
between those who were involved in drawing the picture? about the di erent perspectives re ected in the Rich Picture?
What did you do about that?

System diagrams 35
ff
fl
fi
ff
fi
ff
fl
fl
ffi
ffi
 In uence
Diagrams
There are many di erent kinds of In uence Diagrams. Probably two
of the most widely known are mind maps …

What is an In uence Diagram?

In uence Diagram is a generic name for a group of system diagrams
that provide a snapshot of the variety of in uences operating on a
particular situation.
I use the word 'snapshot' because that's what these diagrams are.
They are freeze-framed pictures of what is in uencing a situation
right now.
In other words, it does this:
An In uence Diagram illustrates the relationships right now,
between factors that are in uencing, right now, a situation right now.
It is a snapshot of a situation, not a video of how a situation arose or
will develop.
This is distinct from a Causal Diagram (see next chapter), which
seeks to illustrate how a situation arose and is likely to change over Source: Guardian https://www.theguardian.com/careers/mind-mapping-tony-buzzan
time.

System diagrams 36
fl
fl
fl
ff
fl
fl
fl
fl
fl
 But they can also illustrate something closer to the de nition of a
… and concept maps: system described in the earlier Systems chapter. Below is an
In uence Diagram of factors that in uence sta retention.

Source: Open University T552 Course

Source: https://www.conceptdraw.com/examples/bubble-diagram-of-internet

Let’s unpick the basic narrative that underpins the development of

In this case a diagram of the various concepts that in uence what we an In uence Diagram.
understand as internet marketing.
Right now I am seeking to explain to you the value and use of
di erent kinds of system diagrams by writing this book. If you
wanted to open up my head and peer into what's going on you will

System diagrams 37
fl
ff
fl
fl
ff
fl
fi
 observe many in uences on why I might want to do that. That
includes a hunch that there is not a general text that covers several
kinds of system diagrams in a practical way, past experience of book
writing that demonstrated that if you want to learn something it is a
good idea to write about it, my experience with facilitating
diagramming exercises and seeing the bene ts of them, the fact that
I have quite a bit of time on my hands because I broke my leg a
month ago when I got tangled up in ski chair lift. Also one of my
clients cancelled the work I was doing because I consequently could
not travel overseas. I also feel inspired to write because I am looking
at the view of the Wellington harbour from my house on a glorious
spring afternoon.
You can probably spot several in uences at play already: My
restricted mobility, the disruption to my work, my wish to share some
experiences and knowledge, my splendid view, my need to keep
myself occupied, the time of year …
So here is a possible In uence Diagram of what is in uencing me
writing this book. The green lines represent in uences that I
consider to be strong in uences, compared with the others. Why draw an In uence Diagram
and not a Rich Picture?
Hang on, I hear you think, what's the di erence between a Rich
Picture and an In uence Diagram? Well, obviously there are no
pictures. But more importantly a Rich Picture seeks to display
everything and anything that might be relevant to a situation that
interests us, such as: the overall knowledge of systemic approaches,
the di erent debates about what does and does not constitute a
systemic approach, the various means of communicating knowledge
other than books, my current physical immobility. Its job is to get our
head around the broad complexity of a situation and how di erent
framings of that situation might alter our understanding of what the
problems may be and what might be involved in addressing those

System diagrams 38
ff
fl
fl
fl
fl
fl
fl
ff
fi
fl
fl
ff
 issues. By contrast, an In uence Diagram tends to be more focused
around a single puzzle within that overall picture; like, why on earth In uence not causality
do I want to spend the next 18 months writing a book on system
diagrams? However, none of the in uence relationships imply causality. I have
found that many people have a problem distinguishing between
Once I have done that, and decided some kind of understanding of
in uence and causality. So here are two examples.
the issue, then I can explore in more detail what might be
in uencing that situation right now, and perhaps suggest some None of the factors I described earlier are causing me to write this
areas that I might need to understand more or change. book. Despite these in uences, I could still end up doing many other
things such as improving my Spanish or doing my physio for the
Yes, the di erence between the two diagramming traditions can
broken leg. Nor will they indicate whether I'll actually complete the
seem subtle in theory. However, in practice there is a di erence
book (although if you are reading this, I presumably did).
between understanding the current situation of knowledge about
systems issues (a suitable theme for a Rich Picture) and Rather, my description was telling you what factors are in uencing
understanding what in uences my decision to sit in front of my me writing the book. If I am not satis ed with one aspect of that
monitors this morning and write these words about system in uence (eg say my knowledge of the potential market for the
diagramming. book), then I may wish to change that situation. But the factors
shown in an in uence diagram do not represent what would require
Once I’ve have identi ed what my main focuses are, in this case,
me to change that situation.
writing a book about systems approaches that include diagrams, I
can diagram how key forces in uence my situation from a particular
perspective at a particular moment in time.
That helps to explain why there isn’t an arrow in the diagram about
how my decision to write a book on system diagrams might impact
on my professional status at a future moment in time. In uence
diagrams illustrate the present situation, the here and now. It is
kind of like the di erence between a snapshot and a video.
Sometimes you just need to know what is the situation right now.
In uence diagrams are always rooted in the moment, interested in
the forces operating here and now, disinterested in those that
operated in past and what might happen in the future – even if they
help me make decisions about the future.

System diagrams 39
fl
fl
fl
fl
fl
ff
fl
ff
fi
fl
fl
fl
fl
fl
fi
ff
fl
fl
 into overcomplicating the diagram.
How to construct an Influence So here is an example of how we decided the focus and scope of an
Diagram in uence diagram.
Recently, I was working on literacy issues in an African country. The
literacy situation in that country is vastly complicated, but relatively
So how do you construct an In uence Diagram? well-known. So it was not immediately obvious where we should
focus. As we talked to people and read the research, it was clear
that students dropping out of school in large numbers a ected the
overall literacy rates. And as we looked more deeply we came
Step One – Decide the focus and scope of the across research that said that one of the three main reasons why
diagram students dropped out was because they disliked school. Yet nobody
had explored the reasons for students’ disliking school. So rather
than focusing on what in uenced literacy broadly, or even what
On the whole, in uence diagramming works best when the process in uenced students to drop out of school, we focused our diagram
is a bit more structured than a Rich Picturing. This can help manage on what in uenced students’ attitudes to their school. And given that
one of the tensions of system diagramming, that is on the one hand, the major drop-outs occurred in the rst three years of schooling, we
having a diagram detailed enough to ensure that the major factors further restricted the boundary of the diagram to primary school
are included, and on the other hand avoiding ending up with a children.
massive wiring diagram that is so complicated you cannot gain any
insights from it.
Compared with Rich Pictures, In uence Diagrams focus more on
what in uences a particular decision (eg ‘write a book’) or state of Step Two – Decide stakeholders and grouping
a airs (eg ‘an environment that encourages book writing’). Your
choice of the boundary and scale will be more speci c. Perhaps just As with all system diagrams it is important to ensure a wide variety of
one of the ‘something to do with …’ framings from a Rich Picture perspectives be applied in constructing an In uence Diagram.
exercise. Perhaps centred on a speci c event or state of a airs. So Probably, the range will be smaller than a Rich Picture. However, you
more than a Rich Picture you need to be a bit more rigid what is can use the method described for stakeholder selection described in
inside and outside the scope of the diagram. In my case it may be the in the chapter on generic issues to consider when working with
aspects that only in uence me to write while I am not able to travel. diagrams. On the next page is part of what we ended up with in the
Or aspects that in uence me writing a speci c kind of book, a self school drop-out example.
published book or one for an established publisher.
When you are working individually you can keep yourself in check,
but groups that re ect many di erent perspectives can easily drift

System diagrams 40
ff
fl
fl
fl
fl
fl
fl
fl
fl
fl
fl
ff
fl
fi
fi
fi
fl
fi
ff
ff
 Involved in The purpose of the Influence Diagram

Stakeholder role Stakeholder role As with all systemic inquiries and explorations, it is useful to be clear
Stakeholder role about purposes. What consequences or results do you want the
in uenced by ... in uences ...
In uence Diagram to contribute towards?
Teacher Curriculum Student knowledge The influences
Pressure from The in uences can be anything that might in uence a state of a airs,
District O cial School curriculum
Ministry event, people or things. For instance:
Village committee Ability of members School curriculum A state of a airs: Feelings, features, resources, organisational
culture, salary levels, students dislike of school, a policy of
Overall attainment of corporal punishment, distance of home from school, season of
Student Enjoyment of school
village literacy the year.
Pressure of A particular event or activity: Book purchase, students attend
Principal performance Culture of learning school, salary negotiation.
measures
People: Teachers, Parents, School inspectors.
Commitment to Things: Type of desk, available teaching technology, the kind
Parent Support at home
education of books in a school library,

In most cases it doesn't matter very much, and the various

Step Three – Decide the Starting Point in uencing factors can be a mixture of states of a airs, events,
people or things. Again, this can be done as part of the stakeholder
analysis, but once you get going you nd other in uences start
Like Rich Picturing, it is possible just to start – once you have
popping up.
decided the aspect of the situation you are wanting to diagram. I nd
it useful to have a speci c starting point or focus for the diagram. In a group session it might be an idea to clarify the di erence
Partly because it is just an easy way to start, but mostly because it is between ‘in uence’ and ‘cause’.
the opportunity for a discussion that might shape the focus and Sometimes I run a short exercise to get people to understand what
content of the in uence diagram. in uenced their interest in, say, attending the diagramming session
There are three sets of decisions to be made before starting. and what caused them to come to the diagramming session. As you
may have already concluded, the distinction between cause and
in uence is a leaky boundary, so be as strict as necessary. What you

System diagrams 41
fl
fl
fl
fl
fl
fl
fl
ffi
fl
ff
fl
fi
fi
fl
ff
fl
ff
ff
fi
 don’t want to include are dubious causal links or have too many ‘it
depends on’ aspects. Step Four – Constructing the diagram
On the other hand, I have learned the hard way that ideas of cause
and in uence may have di erent meanings in di erent languages or
cultural traditions. In fact, in some languages there is no equivalent
di erence. You may have to use di erent concepts such as what
helped you come to the diagram session and what made you come.
Or, say, the di erence between ‘if X happens then Y happens next’
and ‘when Y happened it was helped by …’
And of course, it may not matter a great deal if the two concepts do
get mixed up. It depends on the purpose of the diagramming
process.
But be clear that, as you will see in the next chapter, the conclusions
you can reasonably draw from an In uence Diagram will be di erent
from the conclusions you can draw from a Causal Diagram.
The tone:
Positive, appreciative framing – students like school (state),
children attend school (event).
Negative, de cit framing – students dislike school, children Above is the In uence Diagram we produced using very basic tools.
refuse to go to school. The colour of the ribbons depicted in uence, the colour of the dots
indicated the direction of the in uence.
This is an important decision because an In uence Diagram based
on a de cit framing may be di erent from one based on an Here’s how we did it.
a rmative framing. And the decision depends on what you are trying The process was group based with between four to six people per
to achieve with the diagramming process. In one In uence Diagram group. Normally you aim for groups of mixed stakeholder roles.
we were exploring what in uences the ability of teachers to use their However, in this case there were signi cant power issues so we did
skills in a classroom. We were concerned that if we framed it not mix the groups. Instead, we had a 'promenade' that allowed
negatively then it could promote teacher blaming. So we decided to groups to view and re ect on the diagrams of other groups. The
have the in uencers as things that enabled teachers to use their power inequities were still a factor, but we could manage them more
skills rather than prevented them, and this was re ected in the title of safely.
the diagram.

System diagrams 42
ffi
ff
fl
fi
fl
fi
ff
fl
fl
ff
fl
ff
fl
ff
fl
fl
fi
fl
ff
fl
fl
ff
 In the centre of the paper (two A3 sheets ideal), or whatever medium Once that is complete, the group shares the key in uencers and
being used for the diagram, we wrote down the focus of the decide the most important ve. These can be written on Post-its
In uence Diagram. It should be an end state or outcome, such as (only one item per piece of paper) and surround the focal statement.
children dislike school, or an event such as children refuse to go to Don’t draw any lines yet.
school. Be very aware how the tone of the title can a ect the
The purpose of starting with individual working is to prevent one
discussion and diagram.
person taking over the process. It also slows things down a bit and
Someone took notes of the discussion during the construction of the allows a bit of thought about the in uences. If you have trouble
In uence Diagram. The process of drawing the In uence Diagram is deciding the ve, then use a voting system of some kind.
as important as the diagram itself. It also helps in the analysis.
Try to avoid terms that are too vague (e.g., quality of the school)
Like Rich Picturing, if you are facilitating a group process, then once because di erent people will have di erent ideas what that is. If you
people are under way your main task is to keep people drawing get those kinds of statements, then see they can be broken down
while they are talking. It is not uncommon for people to just talk and into more tangible statements.
not get around to drawing – especially early and later in the process.
Encourage statements that are a direct in uence without any
In fact, many of the issues constructing In uence Diagrams are the important intervening factors. So school books imported from the
same as those with Rich Picturing. So, people will possibly feel USA may in uence children’s dislike of school, not because they are
uncomfortable and uncertain during the process. It can be hard to from the USA but because they use examples that don’t resonate
face other people’s often surprisingly di erent assumptions, because with children based in African countries. And that may in uence
this makes us question our own assumptions. It can mean throwing whether children nd the books boring, which may in uence
away the solutions we thought we had, going back to the beginning whether they nd the lessons boring, which may contribute to their
and starting afresh. This can be demanding and unsettling (and why dislike of school. Another might be ‘weather’ and ‘dislike of school’
a warm-up exercise can be a good idea), but that’s often exactly what where ‘ease of walk to school’ might be an intervening in uence.
is needed at the start of a systemic inquiry. So if you are facilitating
Identify secondary in uences
the process, reassure participants that these feelings are normal
during the process, but it will be worth their while. The group as a whole goes through each of the ve in uences and
I have drawn In uence Diagrams using various processes depending identi es what in uences each of these ve.
on the complexity of the situation. What follows is the most detailed Once these have been agreed by the group, write them on Post-its
version – you can simplify the task by, say, restricting the initial or whatever is being used (again one item per piece of paper) and
number of in uences or the maximum number of in uences. place them around the original ve – close to the relevant in uence.
Identify major in uences Again, don’t draw lines quite yet.
Leave a reasonable amount of time for people to do this. It has to be
Working initially as individuals, identify the most important factors
thoughtfully done. However, allow each in uence to be written. It is
that in uence the core issue. Then identify the three most important.
easier to remove things later than add them.

System diagrams 43
fl
fl
fi
fl
ff
fl
fl
fi
fi
fl
fl
fl
fi
fl
fi
fi
fl
ff
ff
fi
fl
fl
fl
fi
fl
fl
fl
ff
fl
fl
fl
fl
fl
 Like this:
Step Five – Identify the relationships between the
in uences

The group identi es more broadly how the in uences relate to each
other by linking the in uences, generally with lines, with or without
arrows. As with Rich Picturing avoid double heading arrows where
in uences go both ways. Remember that arrows identify directions
of in uence not ows of resources (people, $, things), or sequence in
Rather than this:
time (if X happens then Y happens).
At this point, there is an option to see if there are in uences between
in uences. This can sometimes tidy up the situation where the gap
between in uences are a bit large (such as in the books from the
USA example). Another useful option can be to identify in uences
that are outside the agreed boundary of the diagram. Again, this
needs some time for discussion about whether to include these in
the diagram or that they are things perhaps that are beyond the
ability of the project or participants to do much about. They are in If that happens then take a second look at the diagram and consider
e ect ‘givens’ (and are relevant to the next step). if there are cross linkages.
This is the messy part of the process. If you are working in a group,
there will be disagreements and arguments. This is where pre-
planning is important, since you want to allow people to be able to
move things around and change. On one occasion, exploring the
experience of black and minority college students in the USA, we
were able to use a ten metres long wall that was made of whiteboard
surfacing. So shifting things around and erasing lines was easy. As
you have seen, in the African the relationships were represented by
di erent coloured ribbons, which are less likely to curl up than string,
and a pair of sharp scissors.
Another risk is that the in uence diagram begins to look like a hub
and spoke diagram, where the in uencers radiate outwards, each
in uence just being linked to one other in uence.

System diagrams 44
ff
fl
fl
fl
ff
fl
fl
fl
fi
fl
fl
fl
fl
fl
fl
fl
 Step Six – Identifying the importance between Step Seven – Analyse the diagram and prioritise
in uences responses

Whichever drawing option you choose, once the line drawing is There are many ways of analysing In uence Diagrams. Take a look at
complete, it is critical to have a discussion about the importance and the questions used to re ect on Rich Pictures.
nature of the in uences. The importance of the lines can be signi ed
Also consider the three approaches described in the chapter on
by di erent thicknesses of the lines (see earlier diagram on sta
System diagrams in general:
retention) or the nature of the in uence (e.g., powerful in uence,
political in uence, cultural in uence, motivational in uence) can be Force- eld Analysis
identi ed using di erent thicknesses or colour of the lines, or by Keep, Chuck, Change, Create
writing the information alongside the lines. In the diagram on sta
retention, importance was identi ed by line thickness. Colour is Three Rs
another option, as in the school dropout and book writing examples. In the school drop-out example, we used a mixture of all three.
Finally, to prepare for the analysis stage, more complicated diagrams
can be simpli ed by removing any unnecessary or unimportant lines.
However, it is a good idea to take a photo of the diagram before
starting to remove items.
If this is being done as part of a large group process, where several
In uence Diagrams are being produced, it is a good idea, as in the
Rich Picture example, to start the re ection process with a
promenade, where people walk between di erent diagrams. One
person needs to remain to explain the diagram to people. People
can then return to make any new additions or changes to their own
diagram

System diagrams 45
fl
ff
fi
fi
fl
fi
fl
ff
fl
fl
fl
fi
fl
fl
ff
fl
fl
ff
ff
fi
 Causal
Causal Diagrams and In uence
Diagrams

Diagrams Causal Diagrams may look very similar to In uence Diagrams. The
important di erence is what you wish to know:
If you want to know what in uences a situation at a particular
moment in time then you use an In uence Diagram. You
would use an In uence Diagram if you wanted to know what is
in uencing me writing this chapter, such as the work of other
systems scholars, my experience using system diagrams, my
Purposes of Causal Diagrams idea of what potential readers want to know, what in uenced
the success of my previous books. The core focus is on
features, what may be operating on a particular behaviour.
In Causal Diagrams, the links between factors in the diagram
represent direct ‘if/then’ variable driven relationships with no or If you want to know how a situation changes over a period of
trivial intervening factor variables. time then you use Causal Diagrams. In other words, the
sequence of causes and e ects that in the past have resulted
Causal Diagrams are useful ways of diagramming and gaining in writing this chapter, such as the availability of particular
insights into a situation when its behaviour is either inexplicably software, the demand for a workbook on system diagramming,
stable or things appear to be happening that appear non-linear. Non- the ability to write, the time demands of other work. The core
linear means that the impact of a constant action appears out of focus is about mechanisms, how things change.
proportion to the action itself. A classic example is the performance
of bank accounts where the growth is not proportional to the For instance. Internationally the introduction of non-native species of
individual rate of infection or interest. $100 at 10% annual interest plants into waterways is causing havoc with the local ecology.
means that you have $110 at the end of the year, an additional $10. Several countries have ‘Check, Clean, Dry’ strategies for boats,
The next year at the same interest rate you will have $121, that’s an shing gear and clothing to avoid further spread. If I wanted to know
additional $11 not $10, the following year $12.10. Another is what is in uencing a sher deciding to abide by the instructions, I
pandemics where infections appeared to increase faster everyday, might use an In uence Diagram. I don’t have to worry about
even though the infectiousness of viral strains remain the same. collecting accurate data of cause and e ect, just nd out what is
in uencing that person’s viewpoints and actions. This could inform
An example of inexplicable stability would be a program of my future ideas about the value of the intervention.
vaccination that did not impact infection levels despite its clinical
e ectiveness.

System diagrams 46
fi
ff
fl
fl
fl
ff
fl
fl
fi
ff
fl
fl
ff
fl
fl
fi
fl
 On the other hand, I might want to work out how the rivers and lakes On the left is a Multiple Cause Diagram, on the right is a Causal Loop
got into such a condition in the rst place, and what impact a range Diagram
of strategies (not just Check, Clean, Dry) might have over the next
few years. In that case, I would probably use a Causal Diagram
because I am interested in the mechanisms of spread. But, and it is a
big but, I would have to be very sure that my assumptions around
cause and e ect are accurate, constant and consistent or my
conclusions may be erroneous. This mistake is very common when
assembling and analysing Causal Diagrams and leads to claims that
are potentially incorrect.

Multiple Cause Diagrams and Source: Open University T552

Causal Loop Diagrams Di erences between a Multiple Cause Diagram and a

Causal Loop Diagram
There are dozens of approaches to Causal Diagramming (CD) within
the systems eld and other elds. In terms of method and There are only two major di erences, but they in uence when you
methodologies, it is probably one of the most populated areas of the would use which kind of diagram.
systems eld. I have chosen to focus on two approaches that re ect The factors of a Multiple Cause Diagram are:
two distinct traditions of Causal Diagramming. All Causal Diagrams
have elements of these two, even if they don’t look like them. These • An action (buying more beer)
two traditions go under a variety of di erent names but for the • A state of a airs (loss of inhibition, feeling sick)
purposes of this book, I will call them Multiple Cause Diagrams
(MCD) and Causal Loop Diagrams (CLD). In other words, the factor as stated is either present or absent. If it is
present then that causal chain operates, if it is absent then it does
Opposite is a comparison between the two kinds of diagrams not.
illustrated by the Open University’s System Diagramming Course
T552. The factors of a Causal Loop Diagram are

In this example the situation of interest is how factors a ect beer • Variables; things that can progressively increase or decrease.
consumption. In this case the factor is always present but can change in intensity
and thus a ect the causal chain. It can be considered as a special
case of a Multiple Cause Diagram, but one that adds greater power

System diagrams 47
ff
fi
ff
ff
ff
fi
ff
fi
fi
ff
fl
ff
fl
 to the potential analysis. This chapter covers both kinds of diagram because so much of the
process is the same. Indeed, drawing a Multiple Cause Diagram can
Here is the distinction made by the Open University T552 course
be a good precursor to developing a Causal Loop Diagram.
between Multiple Cause and Causal Loop Diagrams.
The Multiple Cause Diagram will suggest factors that will contain
many potential variables that can be honed down to a single
variable after some further deliberation.
Multiple Cause Diagram Causal Loop Diagram
In addition, some people use both conventions within a single
diagram, although it will limit the causal claims that can be made
Shows a web of interconnected factors compared with a pure CLD.
I will use the phrase Causal Diagram or CD when referring to
Factors something that covers both kinds of diagrams.

The factors are general or The factors are always speci c

speci c events or states variables

Arrows sometimes labelled Arrows labelled with the

with the nature of causality direction of the causality

Helps identify circular loops of Clearly speci es the nature of

causal relationships (aka the feedback loops (positive or
feedback loops) negative, virtuous or vicious)

Useful for exploring rough Adds discipline and clarity

ideas of how changes in because of the need to identify
factors cause changes in other speci c variables that cause
factors those changes

System diagrams 48
fi
fi
fi
fi
 MCD

The features and conventions of

Causal Diagrams
Let me build up a picture of what a Causal Diagram contains by
using an example. In this example the ultimate goal of the
intervention was to reduce the incidence of HIV/AIDS in a mining
town in a west African country. One element of this project was
providing micro-loans to sex workers to purchase hairdressing
equipment. The idea was that the extra income would discourage
the workers accepting more money for unsafe sex. The assumption
was that the availability of loans would enhance the take up of the
micro-loans and thus create more demand for loans.
Micro-loan project MCD
So if we were using MCD based on events or states of a airs, the
key could be: CLD

• Loan giving
• Program reputation
• Loan demand
And would look like the upper diagram opposite.
Using a CLD based on variables, the key factors could be:
• Amount of money loaned
• Degree of support for the program
• Number of loan requests
And would look like the lower diagram opposite.

Micro-loan project CLD

System diagrams 49
ff
Notice the additional convention in the Causal Loop Diagram. The But, in fact, the popularity of the program didn’t behave like that at
arrows are accompanied by a ⊕. This indicates that the variable at all. It oscillated from being wildly popular to wildly unpopular.
the arrow head end responds in the same direction as the changes Like this:
in the variable at the other end. In fact, sometimes the letter S (for
same direction) is used instead of ⊕. As the demand for loans goes
up, so the number of loans goes up. As the number of loans goes up,
so the degree of support for the program goes up. As the degree of
support for the program goes up, so does the demand for loans, thus
closing the loop. Conversely, if the number of loans decreases, the
popularity of the program decreases and the demand for loans
decreases. This is called a positive feedback loop, or a virtuous
circle, or a reinforcing loop – and is indicated by the ⊕ in the
diagram. I prefer to use the term reinforcing loop because the other
two phrases imply that things are always getting bigger, when they
can get bigger or smaller.
So depending on which is happening, a graph of the popularity of Months
the program over time would be a non-linear incremental increase or
decrease shown in these two diagrams. So what might be going on?
Here is one explanation.
Lending impacted on the nancial liability of the agency issuing the
loans. Liability impacted on the kind of controls the lender imposed
on the loans. Those controls can impact on the demand for loans,
which impacts on the loans given.
So we have another loop.

Months Months

System diagrams 50
fi
 An MCD of this new loop would be drawn like this: Notice something di erent here in the CLD. We have three same
direction relationships (blue lines). As the number of loans increases
or decreases, so does the amount of money lent. As lending
increases or decreases so the nancial risk to the lending
organisation increases or decreases in the same direction. As the
nancial risk increases or decreases, the stringency of the loan
conditions increases or decreases in the same direction. But as the
stringency of the loan increases, the demand for loans decreases.
Conversely, as the stringency decreases so the demand for loans
increases.
In other words, the movement is in the opposite direction (green
line). This is shown by the symbol ⊖ or an O (for opposite direction).
The resulting loop is also given various names, a negative feedback
loop or a vicious cycle or a balancing loop. It is called a balancing
loop because it potentially balances a movement in one direction
with a move in another. Balancing loops will always have an odd
number of ⊖ or O (green line) type relationships, because even
numbers of opposite directions cancel each other out. You might
In a CLD it would be drawn in the following way: want to think about that for a moment, but the trick to identify
whether a loop is a balancing loop or a reinforcing loop is to count
the number of negative relationships. An even number (or none at
all) and you have a reinforcing loop and an odd number you have a
balancing loop.

System diagrams 51
fi
ff
fi
So now lets put the two loops together. But that won’t always be the case. Depending on the strength of the
relationship between factors of the diagram you can get all kinds of
patterns, increase, decrease, oscillation, steady state or a mixture of
them all. However, when you get wild oscillations that don’t reduce
in intensity, it indicates that somewhere in the diagram, there is a
delay that is out of proportion to the other parts of the diagram. In
our case example, we found that most of the responses to changes
in one variable a ected the neighbouring variable within a relatively
short space of time. However, we discovered that because of very
poor record keeping and cumbersome decision-making processes,
there was a very long delay between the changes in nancial liability
and altering the loan conditions. In a CLD diagram this is indicated
by two parallel lines. In this possible case, the project managers
spotted this delay and the project funders changed its management
of loans, which brought the program into a balanced, non-oscillating
Now let me describe the cycle. As the project increases in popularity state.
the demand for loans goes up and the number of loans provided
goes up. As the number of loans goes up, so the nancial risk
increases, and so does the stringency of the loans. As the stringency
of loans goes up, the number of loans goes down, as does the
popularity of the program. But here’s the thing, as the popularity of
the program goes down, so does the demand for loans, as the
demand goes down so does the number of loans given, and the
number of loans given goes down so does the nancial liability and
subsequently the stringency of the loans. But as a consequence,
when you lower the stringency of the loans, the demand for loans
increases, leading to an increase in loans given and then an increase
in popularity of the scheme. Hence the oscillations.

System diagrams 52
ff
fi
fi
fi
 So here is the nal CLD Notice how much more speci c the Causal Loop Diagram is. The
factors in the MCD could be represented by many di erent variables,
The nal CLD but because we knew the speci cs of the situation we could pick the
most suitable variables. However, that’s not to say that MCDs are
less useful than CLDs. You need to know much more about a
situation and the speci cs about the relationship between the
variables you chose, otherwise a CLD could end up being less useful
(or reliable) than an MCD. Indeed, I have often seen CLDs drawn on
the basis of evidence that really only supports the drawing of a MCD.
In reality, of course, a CD is likely to be a bit more complicated than
the two easily identi ed loops used in the example. Take this generic
example of a CLD for instance.

In contrast, this is how an MCD might show the same thing.

Multiple Cause Diagram

Source: Open University T552

How many loops do you think there are? In other words, how many
circles can you make by going through a variable only once?
Which ones are reinforcing loops? Which ones are balancing loops?
Recall that balancing loops always have an odd number of di erent
direction relationships.

System diagrams 53
fi
fi
fi
fi
fi
fi
ff
ff
 The ‘non-expert’ framing below focuses on creating strong
Framing the Causal Diagram communities of sustainable consumption at the local level:

Earlier I wrote that how the problem or situation is framed will a ect
the CD and its analysis. And why accommodating multiple
perspectives is important. Here is an excellent example.
On this page are two di erent CLDs of the same identi ed problem
situation but with di erent framings. The situation of interest was
how to bring about sustainable consumption within social and
economic limits in Hungary. One was developed by a group of
‘experts’ and another by a panel of ‘non-experts’ drawn from a local
community.
The expert framing below focuses on the lack of sustainable
consumption at the national level:

Source: Framing Sustainable Consumption in Di erent Ways. Kiss, Pataki, Köves,

Király. https://link.springer.com/article/10.1007/s10603-017-9363-y

Notice that even the leverage points (the red circles) are di erent.
Leverage points are causal factors that signi cantly in uence the
situation. Thus it is critical important to understand that di erent
framings can lead to di erent diagrams and suggest di erent
leverage points.
A nal note – these diagrams are about as complicated visually as
you should get with your diagram. Any more and it is just a swirl of
variables that defeats analysis.

System diagrams 54
fi
ff
ff
ff
ff
fi
fl
fi
ff
ff
ff
ff
 Analysing Causal Diagrams How to construct Causal Diagrams
There are two closely related aspects of Causal Diagram analysis.
One is to generate debate and the other is to generate insights. It is extremely easy to construct meaningless and unanalysable CDs.
Just throw every possible factor against the wall and draw lines
between them and hope for the best. Instead, the approach should
Generate debate be, as with all system diagramming, thoughtful and disciplined about
I always send out late drafts of my books to potential readers for the process. And above all clear about the boundaries, what is in and
comment. A common response was that my micro-loan diagram was what is out.
wrong; that other variables were more important or more revealing. Below is a suggested process, based on a method called Fastbreak
My response to those readers was ‘exactly’. The diagram displayed originally developed by GK Associates. GKA was one of the early
my assessment of the dynamics. It was my causal diagram. Their consultancies that used CLDs.
comments (all di erent, incidentally) indicated another way of
looking at the situation, based on their thoughts, experiences, Brie y it is this:
knowledge and assumptions. In real life, these discussions around 1. Identify the purposes of diagramming
the di erences would help provide more insights into the perceived
problems and solutions. This is another example where system 2. Identify the key framing and the core problem
diagrams are useful tools for generating dialectic and gaining 3. Decide the boundaries of the problem
insights.
4. Identify key factors (events or states for MCDs, and
variables for CLDs)
Generate insights
5. Identify relationships between the key factors – draw
The critical thing to remember (and a common mistake made when CDs
analysing CDs) is that you cannot predict how a particular factor will
behave for two reasons. One is that because you do not know the 6. Identify and categorise important loops and leverage
relative strengths and nature of each relationship. Even if you knew, points for intervention
you could not predict that behaviour because even minimal 7. Validating the diagram (if that ts the purpose of the
oscillations might generate a tipping point. The other reason? There exercise)
is some evidence that the human mind cannot predict behaviour
beyond two or three loops. However, you can gain overall clues
about how a factor is likely to behave and whether it is a major point
of leverage. The diagram can give you an indication of where to look
for what kind of behaviours. Look at CD analyses with a critical eye.

System diagrams 55
fl
ff
ff
fi
 What you will need are:
Step One – Identify the purpose of diagramming
• A group of people who know the issue from a wide range of
perspectives As always, this is the rst place to start. CDs can be constructed for
• Ideally Post-it type hexagons of two colours, or large various purposes. Two common ones are:
rectangular Post-its Explore what may be the key entry points in intervening in a
• A clear wall or oor space – at least 3-4 metres wide problem. In which case validation by comparing the diagram
with ‘reality’ is important.
• Large roll of paper (if you cannot stick things directly on the
wall or oor) Explore how di erent people, stakeholders or whatever
perceive is the dynamics underpinning a problem. In which
• Medium sized permanent markers case, you may want to compare and contrast di erent CDs
constructed by di erent people and gain insights into possible
Planning and involvement di erent problem identi cations and solutions via those
di erent perspectives.
As with all good system diagramming processes, preparation is key.
That is true even when you are drawing the diagram on your own. Decide whether you want to draw a Multiple Cause Diagram or a
But it is especially important when drawing CDs with a group. One Causal Loop Diagram. Some people consider that you should always
reason for this is that CDs, more than any other system diagramming start with a MCD before doing a CLD, because an MCD identi es
process, have an air of certainty about them. While you are not in the general features, that contain many variables. If you start with
prediction business, you will be playing around with scenarios, which variables you might not select the best one.
tend to encourage a mirage of prediction.
So issues about who is involved, what assumptions they bring, how
the process is facilitated, clarity of purpose and su cient time are
critical.
Step Two – Identify the problem focus
I strongly recommend involving a facilitator who not only knows how
to manage groups processes but has a strong background in the However you decide, rst write down the key framing. As described
theory and practice and limitations of CDs. As well, it is a good idea earlier, framing is the perspective or viewpoint through which you
not to rush the early stages of the process (up to and including the want to view the problem and its situation. Is it about health, is it
identi cation of key issues in Step Four). Allow participants to tell about community empowerment, is it about equity, is it about
their stories of their experience within the situation and the issues economic development? Drawing an In uence Diagram or even a
they encountered. Rich Picture may be a good precursor to that decision.

System diagrams 56
ff
ff
fi
fl
fl
ff
ff
fi
fi
fi
fl
ff
ffi
fi
 Framing and expressing the problem will di er depending on which This may appear completely illogical on rst reading. However, the
kind of CD you are using. For instance, let’s assume you are key task of a Causal Diagram is to identify possible places that o er
diagramming a particular situation through a youth justice lens. For a leverage to mitigate a problem. If you have too many items in your
MCD you would want to express the problem in terms of states of diagram, then locating those leverage points will be like trying to nd
a airs or events, like what can be done about our massive youth a needle in a haystack. This is why it is always a good idea to
incarceration problem. That is not su ciently precise for a CLD. explore the issue in detail rst, look at the dynamics of the situation,
Instead, it might be better expressed as what can be done about too nd out what is already known. Start small but don’t let it become too
many young people being convicted of crimes. big. Only include those things that you are sure may be closely
related to that issue and may be an important source of leverage to
Secondly, for MCDs and especially CLDs it is useful to draw a graph
resolving the issue.
of how the key factor (youth incarceration) has changed over time,
What’s the trend? If it is going up or down, what’s the pattern of the If you end up including too many factors (say more than ten to
rise? Don’t get hung up about actual numbers at this stage and don’t fteen), then it is unlikely that the CD is going to bring the insights
try to determine what may be the reason. Either qualitative you hope.
description or quantitative graphing are both perfectly okay as long
But how do you decide what the problem focus should be? It is a
as they are based on reasonably good evidence.
good idea not to rush to the most obvious. For instance, trying to
solve the problem of the #22 bus being consistently late, consider
very carefully before including the entire transport infrastructure for
Step Three – Decide the boundaries the whole city.
On the other hand, if you are trying to solve the problem of all buses
Deciding the boundary of the diagram is critical in Causal Diagrams. being consistently late, or even not turning up (as is the case in my
This is where many problems with CDs start. People try to diagram city right now), then it may be appropriate to include not only the
the whole situation, when in fact you should only diagram those transport infrastructure but maybe other factors, such as school
factors related to the identi ed problems and likely to suggest times, or shopping hours or terms of employment.
possible ways of addressing them. Without this careful deliberation In the micro- nance project there were several important boundary
of the boundary you can easily end up with an unintelligible wiring choices. Time was one where the patterns could only be explored
diagram. over a year long timeframe. The selection of which particular factors
So with CDs you are constructing a diagram of a particular problem, to include or exclude is clearly another. It certainly generated a
not the entire situation within which the problem lies. You only spirited discussion among reviewers of this book’s drafts.
include those aspects of the situation that directly relate to the
problem and potential solutions.

System diagrams 57
fi
fi
ff
fi
fi
fi
ffi
fi
ff
ff
fi
 Step Four – Identify issues
It will look something like this:
Write down the issues that relate to identi ed problems within the
boundaries you’ve chosen. Always be willing to expand or reduce
your boundaries, but explore the consequences if you do.

In a group setting, I generally use the following process.

Each person takes three Post-its and writes down something that
relates to the problem situation. This could be an event, state of
a airs, variable, attitude, behaviour, belief, or knowledge. It could be
something that helps youth incarceration or hinders it. If it is
something really important and powerful then you can use a di erent
coloured Post-it.

Then each person reads out an item in turn and puts up on the wall
or oor or paper in nicely ordered rows. In the illustration opposite
the hexagons have sticky backs that allow them to be stuck to a
vertical surface. You can also use sticky walls using a spray-on
temporary adhesive. Or when neither are available use adhesive
tape. If all else fails, use the oor. Actually, oors are great if there is
su cient space because people can stand around and move things
easily.

Expect about ve to six times as many items as there are people in

the group.

Nicely ordered rows are best because once the process is complete
these individual contributions are numbered starting at 1. This has
nothing to do with priorities; it is easier to number items sequentially
if they are in rows an aids the clustering process – as you will see
later.

System diagrams 58
ff
ffi
fl
fi
fl
fi
fl
ff
 Step Five – Identify clusters of issues

The idea here is to identify the ways in which these issues cluster.
Hexagon shapes tend to cluster better than other shapes such as
squares, pentagons and rectangles.

A fast and remarkably e ective way of clustering is as follows:

Start with a pair of issues that seem to go together and place

them together on another wall, or somewhere distant from the
numbered list. Then repeat for maybe ve or six pairs.

Then open things out a little and just let people cluster ideas
that seem to go together. Make sure that the emerging clusters
are physically separate from each other. The idea behind the
numbering is that people can just say “move #33 to that cluster
over there.” When that is done, the task is to name the clusters. You will
nd that this exposes people’s assumptions about what the
There will be times when an item could go into two or more clusters mean. It is important to allow that debate to happen
existing clusters. Rather than forcing a decision, just put the and allow Post-its to move between clusters and for clusters to
Post-it in one cluster and make another cluster with just the split or merge and be renamed several times. Once everyone
number written on it. That speeds things up a lot. has agreed, take another coloured Post-it, write the name of
the cluster and put it on the cluster.
If you are facilitating, you will nd that after a few minutes
people are starting to name the clusters – just let that emerge
and allow people to change the names as things emerge.

In the end there will be some Post-its that are not clustered.
Just let that be … a cluster of one is okay.

System diagrams 59
fi
ff
fi
fi
 Step Six – Identify CD factors

With luck each of those clusters will represent a state of a airs or

event (MCD) or can be identi ed with one or more variables (CLD)
that can be immediately used to construct your Causal Diagram. In
which case you can ignore this Step and move directly to Step
Seven. If it is not obvious how a CD will develop from the clusters,
then the following questions can help you identify the relevant
factors for your Causal Diagram.

What do key stakeholders want in terms of each cluster?

What aspects of this cluster are known to change?
What are the key choices made (or need to be made) within
each cluster? What helps and hinders those choices?
This process can be done surprisingly fast, even with a large group. What are the key uncertainties?
I’ve clustered 200 ideas in twenty minutes with fty people
participating. On the other hand you can take your time deliberating You may want to discuss this as a whole group, or break into smaller
on each pairing more carefully. cluster groups.

By one of those disturbing truths of human nature, you will nd that If you are drawing a CLD then you need to express the factors as
you will have somewhere between eight and twelve clusters. variables. One way of distinguishing a variable from an event or state
of a airs is that it:
In a large group you will nd there are three kinds of people. There may increase or decrease, and has behaviour that can be
are those who throw themselves actively into the process. There are measured over the period of interest, and
the spectators who stand back and make often useful but occasional
comments. And then there are those who just go and get some measures the intensity or level of something, and may
co ee or check their phones. Allow all three, but make sure uctuate or oscillate.
everyone participates in the debrief, and that the co ee drinkers If you have a list of states or events, then spend some time
bring co ee for the others. discussing what variables might be associated with that state or
event, and then select the most appropriate for the diagramming.

System diagrams 60
fl
ff
ff
ff
fi
fi
fi
ff
ff
fi
 number of negative (O) directions, reinforcing loops will have zero or
Step Seven – Identify relationships between an even number of negative (O) directions.
factors and draw CDs
Some people nd it useful to name the loops. The ‘generating
Now you are ready to start drawing the CD. But remember you need interest’ loop, or the ‘marketing beer’ loop. Keep the names focused
to keep it as simple as possible without being simplistic. Keep in and speci c. A ‘saving the world’ loop is not especially useful even if
mind the problem or focal issue you are seeking to gain insight into. the ambition is laudable. You might even nd the discussion about
what the loop can be called a valuable and insightful task since it will
Identify pairs of events, states of a airs or variables (depending on expose di erent people’s perspectives and meaning-making.
the diagram) that you believe have a direct causal pair relationship
with each other. Again this can result in a very insightful discussion Once you feel you have captured the key dynamics of the loops, it is
about people’s understandings of the dynamics of the situation. time to tell yourselves a story. Start at one of the events, states or
Once you have agreed then write that pair on two Post-its and place variables and walk yourselves through the diagram and while doing
them on the wall together and identify the nature of that relationship. so, tell yourself the story or create a narrative of what you are
Sometimes it is useful to start with physical factors rst ($, people, observing. A bit like I did with the micro- nance CD. You can do this
things). as an individual, group or collective process and then compare and
contrast the stories. Once again, perspectives are important. Each of
Once you have used up your factors in pairs (and there may be some you may see the same thing, but explain it in di erent ways. Discuss
orphans also), start joining up the pairs to demonstrate broader inter- among yourselves if some explanations make more sense than
relationships. This is where a white-board or Post-its can be others. Are any of the stories telling you something you already
extremely useful, so you can move things around for clarity. knew, if so maybe add some things that are uncertain. Does the
story make sense? Can you begin to make some assessments, such
as “oh when X goes up a little bit, Y goes down a lot and that means
…”.
Step Eight – Identify and categorise important Although a diagram is a simple representation of the real situation, it
loops and leverage points for intervention still needs to re ect that situation. To validate the diagram, ask
yourself whether the behaviours described in the narratives re ect
what you actually see over the time period you selected. If not, then
If your diagram is a CLD, then one of the rst things you need to do maybe you need to look again at your diagram.
is to locate your reinforcing loops and balancing loops.

Identify which ones are ‘balancing’ loops and which ones are Finally, identity leverage points. Leverage points are states, events or
‘reinforcing’ loops? Remember that balancing loops will have an odd variables, that if altered can have a major role in resolving the

System diagrams 61
fi
ff
fi
fl
ff
fi
fi
fi
ff
fi
fl
 problematic situation. These can be major clues for identifying
possible interventions.

If you cannot identify any leverage points, then consider either

adding additional variables, or perhaps you need to expand the
boundary of the diagram after all.

For instance, in our micro-loan example, we decided to expand the

diagram to include what a ects the loans being repaid. This actually
helped us identify further leverage points because we discovered
when validating this aspect that in reality the loans were rarely
actually repaid. Many recipients behaved as if they were gifts, while
the lender behaved as if they were loans. So not only did the risks of
lending a ect the terms of the loan, but the actual availability of
money for loans uctuated also because the money available was
meant to be refreshed by loan repayments. This meant that loan
conditions were often decided in a panic when the money was about
to run out, resulting in wild uctuations of program popularity.

But be very careful. Remember that the more things you add the
harder it is to gain insights. Take the diagram opposite for example.
What insights do you gain from this diagram, other than the already
known fact that the situation was complex? In defence of this
diagram, it was developed speci cally to demonstrate that you
cannot tell insightful stories if you include too many variables.

Remember, remember, remember. Diagram the identi ed problem

not the whole situation. And that in general, simple models
provide insights, big complex models do not.

System diagrams 62
ff
fl
ff
fl
fi
fi
 Questions that help analyse and
re ect on Causal Diagrams

Once you have gained a good understanding of your diagram, and

depending on your purposes for constructing a CD, you may nd
these questions useful to pose to your diagram:

Purpose Data
What thoughts do you now have about the purpose of the CD? What kind of data did you use to construct the CD? Who did it
come from?
What data was privileged and what data was excluded or
Scale marginalised?
What conclusions can you draw from the context of the CD? What might be the implications of this for your conclusions?
What was included in the CD and what has been excluded?
What might be any di erences if you expand or reduce the Learning
scale and scope of the CD?

Solutions What have you learned?

What have you con rmed?
What leverage points did you identify? How con dent are you that the assumptions were correct?
What kinds of solutions do these leverage points suggest? What consequence for your conclusions is there if those
What assumptions are you making about those leverage assumptions were incorrect?
points?
And nally ask yourself whether the story the diagram tells enables
Events, States and Variables you to spot where you could intervene to resolve the problem. If you
don’t gain any insights then consider if you have not drawn an
What events, states or variables were excluded from the CD? appropriate diagram.
For what reason, and how appropriate was that?
What assumptions were made about the relative ‘power’ of the
relationships between events, states or variables?

System diagrams 63
fl
fi
fi
fi
ff
fi
 A Coda: Causal Loop Diagrams
and System Dynamic Modelling

There is much confusion about the nature and purpose of Causal

Loop Diagrams and Stock and Flow Diagrams associated with the
System Dynamics approach. Many people often assume they are the
same thing, or that CLDs are a precursor to System Dynamic
Modelling. However, there are two important di erences.

Causal Loop Diagrams are predominantly focused on the

relationship between variables in a situation, how movements
in one variable a ect an adjacent variable. In contrast, System
Dynamic Modelling is concerned with how ’stocks’ (resources
such as people, things, knowledge that can increase or
decrease over time) are a ected as they ‘ ow’ through a
system. In some senses System Dynamic modelling is
simulates the combination of factors that are present in both
MCDs and CLDs

System Dynamic Modelling re ects its name. It simulates the

dynamics of a situation, generally via computer. The Stock and
Flow Diagrams used to express the system’s dynamics are
di cult to understand unless you comprehend the
mathematics that underpin the modelling.

Some say that CLDs are a good precursor to System Dynamic

Modelling, although there is a decades old debate within the System
Dynamic community how useful CLDs are in generating System
Dynamic Models. For these reasons I have not included Stock and
Flow diagrams in this book.

System diagrams 64
ffi
ff
ff
fl
fl
ff
 GENERIC
DIAG AMS

Cyne n

Viable System Model (VSM)

Cultural-Historical Activity Theory

(CHAT)

System diagrams 65
fi
R
 Cyne n
At the time of writing this book, the approach is illustrated by the
diagram below.

Purposes of Cyne n
Cyne n approach is a decision support approach that allows you to
explore, appraise responses to di erent real and perceived
dynamics of a particular situation. Developed by David Snowden,
Cynthia Kurtz and others at IBM’s Institute of Knowledge
Management, the approach identi es ve di erent ways humans
make sense of the world in order to act. It draws not only on systems
and complexity concepts, but also on network theory, learning
theories, and third-generation knowledge management.

Over the years the terminology and understanding of this sense-

making approach has developed considerably.

A situation can be seen in terms of ve domains:

Clear
Complicated
Complex
Chaotic Source: https://cyne n.io/wiki/Cyne n
Confusion (AC in the diagram opposite)

System diagrams 66
fi
fi
fi
fi
fi
ff
fi
fi
fi
ff
 Some of the terminology is not new, the systems literature has used
notions of simple (clear), complicated, complex and chaotic for many The ve Domains described
years. The management literature commonly uses notions of best
practice, good practice and emergent practice as responses to these The Clear Domain
perceived situations.
The domain of the known.
The core features of the Clear Domain are:
If you have come across the Cyne n approach before, you will see
that the terminology has changed over time to better describe
developments in the approach’s conceptual underpinnings. It is a domain of known knowns
Clarity of dynamics. Everyone can see the things the way they
It is very tempting to use this approach as a tool for categorising are. The relationship between cause and e ect is self-evident
situations and the responses to those situations. The situation of to any participant.
serving beer to a customer is simple (clear), a vaccination program is Very simple linear patterns of cause and e ect. If I do X, then Y
complicated, a family is complex, epidemic situations are chaotic. will almost certainly happen.
Indeed, many people use the approach in such a way. From time to
Predictability depends on knowing X and what you are doing
time I do. However, we need to be very careful when we do use it in
with it.
this way because there is much debate whether situations can be
categorised in such de nitive ways. Is delivering beer to a customer The constraints to behaviour, things that can be managed, are
really clear and simple? Very often these categorisations re ect our rigid and known. Behaviour is independent of context.
perceptions of reality rather than be descriptions of reality. And even
There are speci c right answers independent of the current
if they are descriptions of reality, reality is likely to be a mixture of
context.
these states and in constant ux rather than any single xed state.

Example
This chapter covers a variety of ways of using the Cyne n diagram,
but while reading it keep in mind that the approach’s core value is an Take a small school in an isolated rural area in Malaysia. It has been
aid to sense-making. As with all the diagrams in this book, it is not a left money by a wealthy benefactor to provide a library. The library
depiction of reality but a way of helping us understand and respond building needs a concrete oor, a corrugated iron roof, and wood-
to reality. framed walls, exactly the same as all the other buildings in the
village, many of which were built by their occupants. The
construction of the block can thus be perceived as a simple exercise,
building on the known knowledge and expertise within the village.

System diagrams 67
fi
fi
fi
fl
fl
fi
ff
ff
fi
fi
fl
 Complicated Domain books these days. But the answers are out there somewhere – it is a
matter of nding them.

The domain of the knowable. Complex Domain

The core features of the Complicated Domain are: The domain of the unknown
It is the domain of the known unknowns The core features of the Complex Domain are:
The relationship between cause and e ect is not immediately It is the domain of the unknown unknowns
obvious but discoverable through the use of appropriate
Behaviour is highly context dependent.
expertise. Y is predictable if you know X and the details of the
relationships between the two. The constraints to behaviour are enabling rather than
restrictive. They e ectively give people and institutions
The constraints to behaviour are variable but knowable. While
permission to do things.
an individual person or institution may not have all the
answers, answers are accessible via the strong network links The nature of the interrelationships can be determined only
between elements of the situation. during or after the event. X may result in Y or Z or A, but you
will be unable to predict which outcome on the basis of
The constraints to behaviour are restrictive rather than
knowing X or any of the details.
enabling.
The key to understanding what might be going on is by
E ects may be separated from causes in time and space.
observing patterns of behaviour over time rather than single
Expertise is important because knowledge of how to exploit results.
the network is important. Indeed, this domain is sometimes
known as the zone of the expert.
Example
One of the terms of the gift is that the library, although based in the
Example school grounds, will be freely accessible to all members of the
To return to the school library, although members of the village have village and anyone else who is passing through the area. Although
the knowledge and skills to construct the building, there are matters the school has some experience of community use of facilities, the
that require some external expert help. There are building degree of access required under the gift breaks new ground, locally
regulations to be followed as well as conditions laid down by the and indeed elsewhere. There is a rich brew of issues: security, safety,
education authorities. Moreover, local people have very little idea of management, access, nance, control, participation, standards,
what a modern library facility can o er, few have actual experience religious beliefs, and others – all of which interlink in ways that
of libraries, and even they recognise that libraries are more than just nobody can work out in advance.

System diagrams 68
ff
fi
ff
fi
ff
ff
 Cyne n as a tool for managing
Chaotic Domain situations
The core features of the Chaotic Domain are:
Cyne n provides a number of strategies for intervention. I am going
It is the domain of the unknowable to describe three: one focused on interrelationships, one focused
on perspectives, and one focused on boundaries.
Constraints to behaviour are absent or ine ective.
No patterns that can be observed.
Previous experience or knowledge cannot be relied upon.
Interrelationships: Managing E ectively
Example Cyne n suggests four main strategies for managing
interrelationships, each strategy corresponding with the clear,
For instance, the gift is located in a bank that defaults during complicated, complex and chaotic domains.
construction of the building. This was unexpected. What can be
done? Who will pay the outstanding bills? No one knows – at least
immediately – so what can be done?

Domain of Confusion

The Domain of Confusion is an important yet often ignored aspect of

the Cyne n diagram. It describes a state which is not only
completely unstable but also could move into any one of the other
four domains. It can often be confused with the chaotic domain, but
is really a domain of puzzlement and contradictions. It is in fact a
domain where the last diagram based approach in this book (CHAT)
has some relevance.

System diagrams 69
fi
fi
fi
fi
.

ff
ff
 Clear Domain Complicated Domain

The most appropriate ways of managing simple aspects of a In management terms, in uence and a degree of predictability are
situation are the following: possible, while the centre may not know and thus control a situation,
it is able to exploit the strong networks to assess what the
appropriate responses are. Thus:
Sense: Collect su cient data to identify the characteristics of
this aspect of a situation.
Sense: Collect su cient data to identify the characteristics of
this aspect of a situation.
Categorise: Identify where these characteristics t within the
known world.
Analyse: Get the networks to nd out the information and use
Respond: Pick the proven appropriate response to that expertise to evaluate and choose the most appropriate means
category. of response on the basis of that information.

Respond: Pick the most appropriate response to that category.

Thus best practice (context-free application of methods from one
situation to another) is the most e ective, e cient, and e cacious
approach to managing such situations. Thus good practice (context-modi ed application of methods from
one situation to another) is the most e ective, e cient, and
e cacious approach to managing such situations.
Example
To return to the school again:
Example
Sense: What is required is a cement oor, corrugated iron roof, In terms of the design of the library:
and wood frame structure.
Sense: The relationships between the various regulations are
not locally known, but there are experts who can provide
Categorise: There are many examples of this construction in advice. There are many di erent views about what constitutes
the village and a lot of skill around in building them. a modern library but plenty of experience and views about this.
Respond: Errors in the construction are easily identi ed and Analyse: Identify the people and other information sources that
corrected. provide the key advice on how to handle all the regulations.

System diagrams 70
ffi
ffi
ffi
fl
ff
fi
ff
fi
fl
ff
ffi
ffi
fi
fi
ffi
 Respond: Apply this knowledge and expertise to the local
situation. Establish a consultative process about options for the
Example
library. Choose between the various options based on expert
opinion and drawing on other networks of information. In terms of exploring what an open access library means in practice:

Probe: Try two or three strategies or combinations of strategies

such as volunteer sta ng, paid sta ng, pass key access,
Complex Domain timetabled access, access controlled via the local store.

In terms of seeking to manage such situations, it is wise to tread

Sense: Watch the results, spot the patterns that emerge, and
cautiously and to evaluate constantly:
see what works best for whom.
Keep discussion open.
Set clear boundaries within which the elements can operate Respond: Build on what works, avoid what did not, and continue
and self-organise. probing and sensing.
Stimulate attractors to help improve centrality.
Encourage dissent and diversity – they improve the ability to
spot patterns emerging.
Chaotic Domain
Be very careful about starting conditions and monitor for Assert a strong central attractor, hold your breath and see if it does
emergence. any good, and then make an appropriate response to that. Think
triage. Thus:
Thus:
Act: A strong response designed to shock that aspect of the
Probe: Create probes to make the patterns or potential patterns situation back into some form of order, or at the very least
more visible before taking action. triage to staunch the bleeding)
Sense: Collect su cient data to identify the patterns of behaviour
that could be attributable to the probe. Sense: Collect su cient data to identify the patterns of
Respond: On the basis of what seems to be the right thing to do to behaviour that could be attributable to the act.
enhance the patterns that are deemed to be good and dampen
down the patterns that are deemed to be bad. Thus this domain calls Respond: Build on what appears to work, avoid what did not,
for emergent or any behaviour di erent from originally intended. and continue probing and sensing.

System diagrams 71
ffi
ffi
ffi
ff
ffi
 Example
In terms of the bank collapse:
Act: Stop work and pay no bills.
Sense: Observe what happens (e.g., bank gets a bailout,
creditors seek outstanding money).
Respond: Depending on what happens and in which order,
move cautiously.

Perspectives: Multiple Allocations to Domains

But what about the central domain of Confusion?
So far Cyne n has been described in terms of sensing, categorising
or probing di erent aspects of real life behaviour, and responding
accordingly. The underpinning assumption is that everyone involved
in a situation would experience it within the same domain framing. Indeed, frequently, when using the Cyne n approach, di erent
The domain of Confusion, centred in the middle of the diagram, is a people put the same situation according to di erent domains. That
zone where it is not clear which domain an aspect of a situation means they are sense-making di erent ways. If the situation is
resides. problematic in some way, then they may well come up with di erent
ways of addressing that situation. A way of managing this is to allow
all the framings to enter into a dialectic – instead of working out
For instance the diagram opposite depicts a state of confusion or which one is “right,” allowing both to be right. These di erences can
uncertainty what the behaviour of a situation displays and thus be relatively minor such as important exceptions to generalisations,
uncertainty about an appropriate response. important such as situations that can be described from the phrase
‘on the one hand but on the other’, or serious such as unexpected
events or deep puzzles. The solution to any problems associated
with that situation emerge from trying to make sense of the
contradictions that the di erent framings generate.

System diagrams 72
fi
ff
ff
ff
fi
ff
ff
ff
ff
 Example
During a workshop I ran that focused on resolving issues of rural
elder care in Scotland, the same aspect of the situation was placed
in both the complex and the complicated zones by di erent people.
When looking at it, one person turned to another and said, “That’s
why we are disagreeing with each other over how to handle this
issue. You see it as a complex issue and I see it as complicated one.
Let’s talk about that.” In other words, part of the problem they were
experiencing was that they were imagining the same aspect from
two di erent understandings of what was going on. There was an
element of unacknowledged confusion present in the discussions
about what to do. And this, in part, explained why they were having
di culty resolving or managing the situation: “Oh, so you were
managing it as if it were complicated, and I was managing it as if it
were complex – no wonder we were clashing over strategies.” In this between domains. If we can nudge things a small distance across
particular case, the two protagonists were able to come up with a the boundary from one domain to another, we have made the
means of resolving the issue from both perspectives. management strategy easier.

Example
Boundaries: Changing the state of situations After experimenting with various ways of managing access to the
library (perceived as primarily understood within the complex
Managing complex aspects of a situation is resource intensive,
domain), it is found that a pass-key access seems to work better than
complicated aspects somewhat less so, and simple aspects least of
the others, but it does not quite work with visitors to the area. Then
all.
someone recalls reading that a village elsewhere handled access to
If it is possible to ‘nudge’, say, a complex aspect of a situation, or a visitors’ centre by having the pass key stored at the local gas
your perspective of a situation over the boundary into being merely station which was open twenty-four hours. After a bit of research was
complicated, then this is probably more e cient, e ective, and done, the solution seemed to t and was applied – thus allowing that
e cacious than trying to manage it from within a complex domain aspect of the situation to be considered within the complicated
understanding. Can a domain identi ed as complex and thus domain.
requiring being managed using probe-sense-respond, be shifted into
Indeed, the suggested strategy for chaotic situations (act-sense-
the complicated domain and thus managed more easily using sense-
respond) can be seen as pushing the situation into the simple space.
analyse-respond? The general consensus is that this can be
Triage makes a chaotic situation simple, at least for a time and for
achieved when the particular dynamics are close to boundaries

System diagrams 73
ffi
ffi
ff
fi
fi
ffi
ff
ff
 long enough to work out what might actually need to be done. the approach?
A somewhat riskier strategy might be to start asking the kinds of What are the consequences when intervening in that situation?
‘dialectic’ questions associated with the domain of Confusion,
What are the likely impacts of that intervention and how ought
although once you start asking those questions you may not
to be the response?
necessarily end up in the domain of your choice.

Cyne n questions Cyne n resources

This approach continues to evolve. To keep up with the latest ideas
The important thing to remember when addressing situations and
and developments, the best source is the Cyne n Wiki.
problems through a Cyne n lens is that the primary purpose is for
sense-making not description or categorisation. Categorisation is a https://cyne n.io/index.php/Main_Page
means to that end not the end in itself.

Which aspects of the situation can be understood through

which of the ve domains?
Where appropriate:
How can you make sense of those aspects?
How can you categorise those aspects?
How can you probe those aspects?
How can you analyse those aspects?
How ought you respond to any con icting assessments of
those tasks?*
How can the approach help you design strategies for acting or
responding within the situation?
What are your options for acting or responding according to

* See p21 about distinctions between ought and should

System diagrams 74
fi
fi
fi
fi
fi
fl
fi
Viable System
organisations and enterprises. His Viable System Model (VSM)
emerged from these investigations. It is usually displayed as a
diagram – hence its inclusion in this book. The diagram shows that
any human activity system is composed of mutually interlocking and

Model (VSM)
nested Sub-Systems, which when in balance with each other and
with the system’s environment, together provide su cient conditions
for the viability of that system.

The purposes of the Viable 4

System Model

Env
3

iron
2 2

men
t
The Viable System Model (VSM) is a systems approach that identi es 1 1
the minimum requirements that must be placed on collective human
1 1
endeavours if they are to prove enduring and capable of
development. The aim is to achieve optimum performance in terms
1 1
of the systems’ purposes and able to adapt and survive in a complex
and dynamic environment. In other words, be viable.
System 1…operative units
System 2…coordination and information
System 3…management and control
System 4…development and interaction with the environment

The VSM diagram explained System 5…policy, coordinates internal and external, present and future

In the 1970s, the British cyberneticist Sta ord Beer explored the Source: Adapted from Systems Concepts in Action. Williams & Hummelbrunner.
viability conditions of living systems, such as biological organisms, to Stanford Press (2010)
see if they could relate to socio-cultural systems, such as

System diagrams 75
ff
ffi
fi
 The primary focus when seeking to achieve viability is how well Patrick Hoverstadt (see VSM resources for reference) identi es four
variety is managed with the system, and between the system and its main drivers of complexity in Sub-System One that collective pose
environments. I nd the term variety a bit confusing because in variety challenges:
everyday speech it has a fairly broad meaning. Its meaning in VSM
How many di erent things are done within the Sub-System?
draws on Russ Ashby’s concept of requisite variety. Variety is the
total number of states that a component of a system can possess. If How many di erent places things are done in?
two components that interact with each other have the same number When and which things are done?
of potential states, then the relationship is said to have requisite
variety, the two components are in balance and thus the system is For whom or what they are done?
viable or sustainable. Indeed, Ashby called this concept the Law of Clearly in complex and dynamic situations there will always be a
Requisite Variety. I have always found that this explanation is rather tension between how to manage too much variety or too little variety
abstract, but in simple terms it means something like this. to achieve requisite variety. The variety of individual parts of a
If a call centre has 20 lines, and that at any one time it has 20 system need to be controlled to ensure that the entire system has
enquiries, then that call centre has ‘requisite variety’. But if that requisite variety and is thus viable. In other words, the call centre is
demand varies and has a peak of 60 calls at any one time, merely part of a larger system. It contributes to the variety of the
then it does not have requisite variety to respond to the entire system and has to adjust its variety in order to stabilise the
demand, since the variety of the centre (20 simultaneous system as a whole. VSM was developed to help address exactly that
conversations) does not match the variety of its environment complex set of decisions.
(which has a peak of 60 needed conversations). If that situation The VSM diagram highlights ve Sub-Systems that can be observed
continues the call centre will not be viable long term because it in every human activity system. The management of these tensions
will not satisfy the needs of its environment. within and between them are critical.

Obviously that is just a simple example. Every component of a Sub-System One: Operational activities
situation has many dimensions of variety. The call centre can be Sub-System Two: Coordination and information
large, or small (size variety). The sta can be paid a salary or on the
number of customers they handle (employment type variety). They Sub-System Three: Management and control
can have a narrow or wide range of authority to directly address Sub-System Four: Development and interaction with the
issues raised by callers (devolution variety). And these varieties environment
themselves may change (dynamic variety).
Sub-System Five: Strategy, identity and sustainability
As Hoverstadt points out, the core task of managing these tensions
generally boils down to managing the tension between autonomy
and control.

System diagrams 76
ff
ff
fi
fi
ff
fi
 Particular parts of the system need to have freedom to respond to
their own needs for variety (e.g., the call centre can have as many The VSM analytical process
lines as it needs to respond to the environment’s demands). If any
component in a system has too much autonomy then it could Let’s look at each of the ve Sub-Systems of VSM and then pose
possess so much variety that the other components cannot handle it. some questions relevant to viability issues.
On the other hand if there is too much control – low variety – each
Before I do this, I want to highlight a potential trap that Patrick
component may not be able to respond e ectively to changes in
Hoverstadt identi es about how to think about these ve Sub-
those parts of the system and environment it interacts with.
Systems. The ve Sub-Systems could be seen as forming a
hierarchy. Thus Sub-System Five controlling Sub-System Four, which
Notice the parts of the environment comment. This highlights a controls Sub-System Three and so on. In fact, it is much more
common problem within many systems approaches. They assume accurate to view them as ve interrelated sets of roles,
that the environment is a single undi erentiated blob and that a responsibilities and actions that need to be kept in su cient balance
system interacts with it also as a single undi erentiated blob. As the in order to manage (individually and collectively) the ever-changing
VSM diagram shows, an advantage with VSM is it acknowledges that needs and conditions of the world around them. No Sub-System is
di erent parts of a system will interact with di erent parts of the any more important than the another, and all contribute to the trade-
environment. Of course this raises variety issues with the o s necessary for managing a viable endeavour.
environment that need to be managed if the system is to be viable.
Unfortunately, the way this chapter describes these Sub-Systems
could promote that hierarchical view. However, that is because (as I
mentioned right at the beginning), writing is linear whereas a
diagram is not. In a real analysis you would keep looping back and
forth as the answers to one set of Sub-System questions a ect and
are a ected by the answers to another. For instance, the answer to
the rst question on Sub-System One to some extent depends on
the answers to Sub-System Five. But it is di cult to answer Sub-
System Five questions without some assistance from the answers to
Sub-Systems One to Four. Iteration is key to the success of a VSM
inquiry.
What’s more, if we return to the systems triangle above, you can see
that not only will there be real di erences in variety within the
activity system and with the environment, but there will also be Having issued these cautions, let’s consider Sub-System One.
perceived di erences in variety. Indeed, those di erent perceptions
themselves contribute further to the variety of the system. Both real
and perceived varieties need to be managed in order to sustain
system viability.

System diagrams 77
ff
ff
fi
ff
ff
fi
fi
fi
fi
ff
ff
ff
ffi
ff
ff
ff
fi
ffi
ff
 Sub-System One: Operational Sub-System Financial support to those seeking work and those providing
work opportunities.
Sub-System One is essentially what the system does to satisfy the
immediate demands of the external environment. In other words, the Questions for exploring Sub-System One (Service Delivery)
delivery of front-line services. Within any System there will be several
Sub-Systems Ones, although often one of them is regarded as the VSM can be used for diagnostic purposes (in an is mode) or design
primary Sub-System, with the others in a supporting role. However, purposes (in an ought mode). For the sake of clarity I have phrased
selecting which one is primary depends on how the system is everything in the design oriented ought mode, but the questions
framed. Thus a needle exchange scheme – where injecting drug work equally well in a diagnostic is mode.*
users can swap a used needle for an unused one – will contain many
What ought to be the primary Sub-System?
operational activities such as publicity, needle provision, needle
acquisition, medical care, community relations, security. However, if it How can we ensure that the primary Sub-System has su cient
is framed as a health system, then the primary activity relates to the variety to respond to the variety required by the environment
sourcing and provision of clean needles. However, if it is framed as a and the system’s intended bene ts? In particular, in terms of
safety system then the primary activities will relate to ensuring that technology, geography, time and bene ciaries?
drug users visiting the venue will be able to access, remain in and What would be the same or di erent with a di erent framing
leave the venue without being harassed – the security activity. that changed the primary activity?
Disability services example
It is a massive challenge to deliver services to people with complex
needs that cross organisational boundaries. The primary example Sub-System Two: Coordination and information
used in this chapter is based on a VSM inquiry into how a range of
governmental and non-governmental agencies, with di erent policy,
governance and delivery roles, enabled (or not) access by people In highly complex human endeavours (such as health or education
with disabilities to work opportunities. services), the demands of each individual or stakeholder role in the
environment will di er. There will also be many di erent
The inquiry identi ed many variety problems in components of Sub- perspectives or framings in that environment about the purpose of
System One, including: that endeavour. Many operational units will be involved, requiring
many di erent levels of variety.
Equality of access to services that were supposed to enhance
job opportunities. Also, given the range of operational (Sub-System One) activities it
Availability of information about job opportunities for those takes to produce the desired bene t to the environment, it is vital
seeking work. that not only does information pass between these Sub-Systems, but

* See p21 about distinctions between ought and should

System diagrams 78
ff
fi
ff

ff
fi
fi
fi
ff
ff
ff
ffi
 that the variety of that information supplied to each Sub-System has Disability services example
to match the variety of the information required by those Sub-
Systems. Variety issues around co-ordination within and between agencies
included:
In other words, information has to have requisite variety so that the Too much information variety. For instance, the same
operational Sub-Systems are not starved of, or overwhelmed by, information was required from individual people with
information. And because coordination implies an exchange, disabilities to multiple service providers, duplication of
information requirements by one Sub-System from another has to information between providers.
match with the other Sub-Systems’ capacities to supply it. It should
Inadequate variety of service provision in some areas and too
be clear by now that in real life there will always be trade-o s about
little in others leading to under-serving in one area and
the demand for and supply of information in order to get as close as
duplication of services in another.
possible to requisite variety.
Reporting processes were vertical rather than horizontal. So,
No surprises then that Sub-System Two is often one of the most for the providers of local services to people with disabilities,
critical aspects of any VSM inquiry. It plays a vital role coordinating the information ow was largely to national funders and
the various Sub-System One units by bringing awareness and centralised management providing a barrier for coordination
enabling responses to choice and variety from the environment, between local service providers.
such as ensuring that the call centre workers know how to lter out
irrelevant calls. It also has to coordinate the implementation of
decisions made by other Sub-Systems, for example preventing call Questions for exploring Sub-System Two
centre workers from making promises that are against the interests
How many distinctive yet interdependent operational units are
of the system. It also has to enable boundary choices to be
necessary to provide the desired bene ts in the environment?
sustained – whether a school physics department is teaching the
exactly the same thing as the chemistry department and thus How can the variety be matched between these units,
possibly generating surplus variety. between these units and Sub-Systems Three to Five, and with
essential exchanges with the environment?
I recall talking to a call centre worker of my cellular company about
How ought the units share their knowledge of sudden
my inability to access the service I was paying for. I was told that the
changes in the environment?
sales and promotional activities (Marketing Sub-System One)
operated independently from the engineers, equipment purchasers What kind of information and coordination issues repeatedly
and software developers (Procurement and Technical Sub-System arise – and how much has that to do with inadequately
One). Consequently, services were being marketed that were not yet matched variety? What ought to be the response to that?
ready to release. Too little variety within Sub-System Two led to too How ought boundary issues between operational units be
much variety in Sub-System One. identi ed? Boundary issues can relate to time, resources,

System diagrams 79
fi
fl
fi
fi
ff
 standards, knowledge, autonomy and control? Disability services example
How good is the information about the information? Do The management of the services was very hierarchical and did not
individual units understand the importance and role of always involve those delivering the services on the ground in its
information they provide to other parts of the System? decision-making around resource allocation, standards of
performance and evaluation. This led to resource allocation and
project funding issues that resulted in potential bene ciaries either
receiving services of varying quality (too much variety in standards)
or gaps in service provision (too little resource variety). The
Sub-System Three – Management and Control monitoring processes were largely focused on accountability
It is easy to con ate Sub-System Two and Sub-System Three. After mechanisms within individual agencies, rather than being focused on
all, isn’t the role of Management to coordinate? But this confuses the quality of results to supposed bene ciaries or the e ectiveness
Management (an entity) with management (a process). Sub-System of inter-organisational coordination. Consequently, there was a mis-
Three is about management of resources in the broadest sense, not match between the variety needed at Sub-System One level to
an organisational tier of activity. Essentially, Sub-System Three is the service its environment and the variety provided to it by Sub-System
Sub-System that resources, establishes, monitors, evaluates and Three.
sustains the rules, standards and procedures by which Sub-System
Two assists the development of appropriate variety of Sub-System
One. It ensures that the various components of Sub-System One and Questions for exploring Sub-System Three
Sub-System Two and the relationship between those parts is not What ought to be the rules, standards and procedures by
only coordinated and informed but importantly creates a synergy which Sub-System Two in uences and maintains the requisite
that satis es the demands of the customer, client or intended variety of Sub-System One?
bene ciary of the system. In other words, if Sub-System Two enables
Sub-System One to do its job, Sub-System Three manages the How ought these be measured and monitored/evaluated?
resources for Sub-System One and Sub-Systems Two to do their
jobs, and checks on their performance. It also provides a link
between the more operationally oriented Sub-Systems One and Two
and the more strategically oriented Sub-Systems Four and Five.
One of Sub-Systems Three’s most important functions therefore is to
develop and manage a way of measuring the performance of the
entire system. An important aspect is that measurement focuses on
whether the system is actually providing the intended bene t and
the most appropriate trade-o s of variety so that the bene ciaries of
the System are in fact getting the intended bene ts.

System diagrams 80
fi
fi
fl
fl
ff
fi
fi
fi
ff
fi
fi
 Sub-System Four: Development and interaction with Sub-System Four did not identify that the strategic policy cycle (see
Sub-System Five), was too slow and out of synch with the rapid
the environment development of employment sectors in the wider labour market that
provided potential opportunities for people with physical disabilities
This is another potential area for misunderstanding. All the Sub- such as gaming and IT support services.
Systems have responsibilities for keeping an eye on, and managing On the other hand it was found that reasonable accommodation
variety with the environment. Sub-System One scans the policies were becoming a focus for more stringent monitoring, and
environment to ensure that it ful ls customer (bene ciary) needs. similarly that initiatives were being developed to promote self-
Sub-System Two has to ensure that the coordination functions match employment and entrepreneurship were also very sensitive to the
those expected by the environment. Sub-System Three will monitor needs of those with disabilities.
the issues relating to trends in issues like international standards and
practices.
Questions for exploring Sub-System Four
So what is special about Sub-System Four? Sub-System Four has the
responsibility for developing, engaging and monitoring the How ought the external environment be scanned for changes
environment as a whole so that the entire System has su cient or potential future changes – speci cally to identify variety
variety to cope with the variety of the existing environment and risks?
prepare the system for anticipated or unexpected changes in the What level and kind of external communications are required
environment. Part of Sub-System’s Four role is to assess the present, to adequately scan the environment?
part of its role is to anticipate the future. In other words, Sub-System
How ought the communications within the System be
Four is responsible for designing and resolving issues raised by
managed so that it doesn’t add too much variety to other
change processes to enable the System to respond to changes in its
System components?
environment. It thus needs to be well integrated into activities and
variety issues in all the other Sub-Systems. Is the System actually providing the intended bene ts by
checking on Sub-System One? Are those bene ts satisfying
the most important bene ciary needs to the right
Disability services example bene ciaries?
Variety problems were identi ed around balancing tensions between How aware is Sub-System Four of possible boundary
a rights-based approach to disability and employment – equity of problems? For instance, does the system have legitimacy?
access to work by people with disabilities and the economic
imperative for growth and access to a skilled workforce. In practice, Are certain parts of the environment being excluded or
insu cient attention was given to ensuring that people with marginalised for practical, cultural or political reasons?
disabilities have access to decent work that adequately addressed Examples could be trade unions, organisations perceived as
in-work poverty. too radical, or limited framing of what ‘the problem’ is?

System diagrams 81
ffi
fi
fi
fi
fi
fi
fi
fi
fi
ffi
 Sub-System Five: Strategy, identity and sustainability relationship with the other four Sub-Systems. E ectively, it has to
manage the entire network of variety trade-o s so that as a whole
the System and its environment share similar varieties, now and in
Sub-System Five is often described as being responsible for the future. In the end it bears prime responsibility for balancing the
addressing the existential questions, who are we and why should we necessary variety to allow autonomy in the right time and place, and
exist? It establishes policies in light of competing demands between control in the right time and place. It has to work out how to keep
the present and future and between internal and external action tight enough to avoid disintegration of the System, but loose
perspectives. If Sub-System Four is about identifying the present and enough to adapt to variety changes within the System and within the
predicting the future, Sub-System Five is about helping create the environment.
future system while still maintaining the viability of the current
System. Disability services example
However, this assumes that it is obvious or self-evident who or what There were several Sub-System Five variety issues. The governance
the ‘we’ is. Who or what is in the System and who or what is in the arrangements for the System were too restrictive relative to
environment? In the current gig economy era what does mainstream employment and disability policy. Also the agencies
employment mean? What is understood by disability? Are the non- responsible for disability polices were divorced from, knew little
disability aspects of NGOs that provide a range of services to many about and unable to in uence wider employment service policy and
di erent types of clients part of this System or not – especially if practice. There was a signi cant variety imbalance between
those aspects provide valuable cash ow and information that may opportunity and operation.
be useful to this Disability System? At the broadest strategic level the Sub-System Five failed to
acknowledge that di erent parts of the employment landscape were
Sub-System Five makes those choices. working to di erent purposes. There were competing purposes;
short term political agendas, the drive for economic growth and the
In other words, we are back to the triangle diagram depicting the human rights agenda. So essentially, the variety was there out in the
balancing of interrelationships, perspectives and boundaries. Sub- real world, but the variety of the agencies to respond to that degree
System Five has decide what ‘The System’ actually is and what of variety was too small. This returns to an earlier comment that
ought to be the key boundary choices that determine what that perceptions of variety are themselves a form of variety.
System is. And having done that, make decisions about what’s in this
System and what is out of this System. And what to do about the As identi ed in Sub-System Four, there was an absence of a shared
di erent perspectives that will be privileged or ignored by these agenda within the System around the development of policies that
decisions. centre on preparing people for the future of work (insu cient
variety), and in relation to Sub-Systems One to Three signi cant
Although Sub-System Five bears the primary responsibility for doing imbalances between specialist support services preparing people
all of this, clearly it can only do so in conjunction with a deep mutual with disabilities for work and agencies providing work opportunities.

System diagrams 82
ff
ff
fi
ff
ff
fl
fi
fl
ff
ff
ffi
fi
 Operational reviews (see Sub-System Three) were narrowly focused variety of information for each Sub-System so that they can
– low variety – and unlikely to produce the level of reform required e ectively communicate with each other?
to create the level of exibility and adaptability to respond to the
changing nature of the workplace (Sub-System Four). Focusing on
access to and performance within institutions, rather than And nally, most importantly:
transformation, (Sub-System Three) did not result in inclusion of What ought to be the overall balance between autonomy and
greater number of people with disability in employment (Sub-System control?
One).

Questions for exploring Sub-System Five VSM as a re ective process

How ought the System be framed? How many di erent ways
can you complete the phrase ‘This System is something to do VSM was designed to investigate entire interventions, initiatives and
with …’ institutions. However, colleague have also used VSM as a way of
What ought to be the Systems purposes (plural) and who or structuring personal re ection on the roles and practices in their own
what ought to be the bene ciaries for each purpose? work. The kinds of questions posed include:

What ought to be the System’s image of itself? When you say you work in Sub-System X and Y, what does that
mean in practice?
What ought the System do to gain and sustain legitimacy from
the environment and respond to the variety requirements of If you feel you should work in more in Sub-System Z what do
the environment? you anticipate it would look like?

How ought Sub-System Five maintain the balance between Would any shift move you towards achieving requisite variety,
Sub-System Three and Sub-System Four in formulating or overload you or the System with too much variety? If the
strategy? latter, where is the sweet spot? For whom? Under what
circumstances?
If, as a result of this scanning, there is a need to change the
System, how ought that be done so that it doesn’t add too
much variety to the system as a whole? What ought to be the
balance between maintaining the status quo and development
(the ying the plane while building it conundrum)?
How should Sub-Systems One to Four exchange information
with each other, that both satis es the needs of each speci c
Sub-System? In other words, what ought to be the requisite

System diagrams 83
ff
fl
fi
fl
fl
fl
fi
fi
ff
fi
 There is, of course, the risk that a Viable System based inquiry can
Recursiveness just go on for ever, diving into progressively smaller sub-Sub-
Systems. In practice, for me it is about usefulness. If it is useful to
Sta ord Beer regarded the VSM diagram as depicting a recursive bring in the recursiveness concept, then I do so. If I judge that it is
system. He used this word to describe the fractal nature of a viable not important relative to the situation or problem I am trying to
system. In plain language, no matter how many times you divide the resolve then I don’t. It is a boundary choice. Like every boundary
System up into smaller units, the basic ve Sub-Systems still exist. So choice it is a mixture of pragmatics and addressing the implications.
each component of Sub-System One contains all ve Sub-Systems,
as does each component of Sub-System Three.
Think of these:
Viable System Model (VSM)
resources
VSM is a subtle approach with many di erent ways of approaching it.
There is a lot of material on the Internet, much of it highly technical.
There are also several variations of the basic model. I cannot
suggest any overall website reference. However, I strongly
recommend the chapter on the Viable System Model by Patrick
Hoverstadt in the following publication:
M. Reynolds, S. Holwell (eds.), Systems Approaches to Making
Change: A Practical Guide
This recursive structure ensures that down to the smallest viable
https://doi.org/10.1007/978-1-4471-7472-1_3
component, each component can be seen to have a policy,
intelligence/information, control and coordination functions that The inspiration for the case example was O'Donnell, J. and Schmidt,
enable the component to be viable. For instance, this component B. (2021) Integration of people with disabilities into the open labour
makes localised policies about how to handle the problems they market: Irish case study. Eurofound Working Paper WPEF21044.
encounter in their day-to-day tasks and provides the exibility to However, the interpretations are mine, and in some cases I changed
survive in complex and rapidly changing environments (Sub-System the story a little to clarify a point.
Five). Recursiveness permits some degree of autonomy to operate
https://www.eurofound.europa.eu/sites/default/ les/wpef21044.pdf
within the necessary degrees of control.
DOI: 10.13140/RG.2.2.18754.50880

System diagrams 84
ff
ff
fi
fi
fi
fl
 Cultural-
Historical Activity
eory (CHAT) Finally we get to Cultural-Historical Activity Theory (often shorted to
Activity Theory, Activity Systems or CHAT). It is an exceptionally
detailed systems approach and it is easy to get lost in that detail. I
have included this approach because it is detailed and the diagrams
associated with it are an excellent example of how to explain a
particular systems approach through diagrams.

Tools and The approach supports a seven-stage analytical process, where

Artifacts each step poses questions that reveal relevant systemic issues. It
demonstrates the power of a diagram as a vehicle to reveal systemic
mediate questions.
People Environment Needs These diagrams provide a means not only of helping to get your
acting on to satisfy head around the CHAT approach as a whole but as a useful visual
reminder of each component and its associated questions.
Rules Community Organisation Nevertheless, like VSM it is a complex approach. Because of this
Formal Those around Work practices/roles complexity, I really encourage you to step through the explanations
I nformal us and diagrams with a speci c project or issue in mind. Otherwise you
(cul ture) can get easily overwhelmed by the detail. Using an example will
provide the necessary stepping stones to ease your understanding.
Source: All diagrams adapted from Activity Theory As A Design Principle For Team So take your time with learning this approach. It will be worth it.
Development Processes. Capper, Hill, Wilson, Harvey. www.bobwilliams.co.nz/
ewExternalFiles/CHAT.pdf

System diagrams 85
Th
fi
 Purposes of CHAT What is CHAT?
The core purpose of the CHAT approach is to enable you to promote CHAT is based on established systems principles but takes a
creativity and learning in collaborative working. It is especially useful radically di erent approach to them.
when that creativity and learning is necessary under challenging
The most common system approaches were developed by
circumstances. Such as:
biologists, engineers and sociologists whose starting point was on
People with di erent roles, values, skills, histories, cultures and inter-relationships and then later incorporated how di erent
goals having to focus on a common activity. perspectives a ect observations of those interrelationships. By
contrast, CHAT was developed by cognitive psychologists and
Sudden changes in the system, such as new policies, or in its
organisational developers. Their starting point was perspectives and
context, such as new legislation, disruption due to an extreme
how they in uence our understandings of the real world, draw
event.
meanings from that understanding, create learnings from those
Contradictions within the system. For instance, keeping a meanings and are motivated to respond on the basis of those
workplace safe and pro table. learnings.
Thus, CHAT challenges common systems approaches in the
following way. If the point of systems thought is to gain new insights
CHAT history and meanings, then shouldn’t systems approaches incorporate
theories of learning? If the way we learn from the real world is
essentially a cognitive process, then the systems approaches we
CHAT was originally developed by the historical philosopher Alexei develop should be based less on what is happening in the real world
Leont'ev, at Moscow State Lomonosov University during the 1960s, and more on the cognitive processes involved when observing the
based largely on Lev Vygotsky’s particular ideas of cognition and real world.
learning. Activity Theory was subsequently the basis of much
research in Russia, especially in the areas of play and learning. A CHAT based inquiry combines three components:

In more recent years the implications of Cultural-Historical Activity A systems component – that helps you to construct meanings
Theory in organisational development have been promoted by the from situations.
work of Yrjö Engeström’s team at the Centre for Activity Theory and A learning component – a method of learning from those
Developmental Work Research at the University of Helsinki, and meanings based on Vygotskian theories of developmental
Mike Cole at the Laboratory of Comparative Human Cognition at the learning.
University of California San Diego campus.
A developmental component – that allows you to expand
those learnings towards action.

System diagrams 86
ff
fl
ff
ff
fi
ff
 Constructing the CHAT diagram Proposition Two – Activity
‘Activity’ is what happens when human beings operate on their
These three components are constructed from seven basic environment in order to satisfy a needs state.
propositions. It is important to note that unlike most of the systems eld, CHAT
The rst ve propositions comprise the systems component uses the word environment to mean the context within which people
operate. It describes components inside the system boundary not
The sixth comprises the learning component
outside the system boundary. Common-place language does the
The seventh comprises the developmental component same. We often refer to the work environment, the school
environment, a learning environment.
Out of each proposition ows a set of systemic questions.

People Environment Needs

acting on to satisfy
Proposition One – The Fundamental Proposition
Activity Theory is based on the proposition that learning is a social
and cultural process not simply a cognitive biological process. The needs we are seeking to satisfy provide the motive for the
activity more than do the actions we are undertaking as part of that
The proposition means that in di erent situations thinking and activity. This is because individual actions may have di erent
learning will be practised and achieved in di erent ways. motivations. For example, a horticultural system needs produce to be
Furthermore, those ways are not likely to be transferred easily or picked on schedule. However, a farm supervisor seen to be
directly from one person, team or organisation to another. It thus pressuring the team of orange pickers to work faster might be
challenges most of the common ideas of how you learn, via books, individually motivated more by looking good so that she can be
lectures, teaching or whatever. You may be aware of the concept promoted to a desk job. The fruit pickers are individually motivated
‘communities of practice’ – CHAT was the systems approach from to pick faster so that they can earn enough money to send some
which the concept emerged. back to their families in Guatemala. But initially, at least, all these
motivations are compatible with the collective motivation. We shall
see later when that is not the case.
Similarly there are many di erent actions we might take to satisfy the
same need. For example, one farmer might try to become more
innovative by reading books, while another might set up a discussion

System diagrams 87
fi
fi
fl
ff
ff
ff
fi
ff
 group with other farmers to achieve the same end. highlighted the di culties of the system and Local Stu within it
surviving as a socially and environmentally conscious business in a
highly competitive context. Within the activity system, Local Stu in
Proposition Two based questions particular has had to be highly innovative and totally committed to
learning for the activity system to have survived. In particular, it has
What can you observe happening?
had to respond to sudden shocks (cancellation of the supermarket
What needs are those actions supposed to ful l? contract), major contradictions (supply of non-local produce), and
rapid changes in roles, such as shifting from a delivery focus, to a
Do the actions within the activity appear to t these needs? Or
retail focus to a consultancy focus.
might there be some other, unstated, tacit needs that are
determining the actions?
Proposition Three – Information
How well suited are the actions you can observe to the stated
needs? If there is a mis t, at a rst guess why do you think that Information must ow through the activity system in order for the
is happening? desired needs to be ful lled.

Example – Horticultural Activity System

Information about
the environment
Local Stu is a grocery business in Vermont USA with the primary
aim of promoting the availability of locally produced food to bene t
small producers, workers and to reduce carbon footprint . It has People Environment Needs
been around for many years and has evolved through di erent acting on to satisfy
stages, from a small business to a medium sized company that has a
Information about
sales arm and a consultancy arm. It is supported by income from
our needs
sales, consultancies as well as grants from philanthropic foundations
and government agencies to support some of its services. For
instance, it employs people with learning di culties. So the activity
There are two types of information.
system here is considered to be the farm based producers, the retail
based customers with Local Stu as an intermediary. It has been a 1. Information about our needs (What are they? Have they
bumpy ride. Local Stu itself is committed to sustaining excellent been satis ed?) comes from within ourselves (and others
relationships with growers and retailers, a situation which has been who are part of the activity system) and involves an internal
tested by rapid expansion, risky investment in growth strategies and dialogue, either personally or interpersonally.
over-reliance on some large non-local suppliers (for out-of-season 2. Information about the environment What resources are
fruit) and a large single supermarket chain. These conditions have there that can help me/us/them meet those needs? What

System diagrams 88
ff
fi
ffi
fl
ff
fi
fi
fi
ff
fi
ffi
fi
ff
ff
ff
fi
 has to be done with those resources in order to succeed? vitally important to performance. Yet often, for practical reasons,
This comes from outside via our senses, seeing, hearing or many individuals, teams, program participants get their information
feeling. However, both types of information require internal unequally, and may have di erent internal resources and
mental processing to make sense of them. perceptions for making sense of the information.
But there are also two main types of information we get from the
To put it colloquially, the data may be there for all to see, but it is
environment, and they subdivide. The main categories of information
spoken about in many tongues. Making sense, collectively and
we receive looks like this:
individually, of what the data means is a complex, emergent process.
But, fundamentally, when we decide ‘what to do’, we place the
information we have gathered alongside our needs and gure out
The physical Non verbal Spoken Written how to use one to satisfy the other. And that raises the issue of who
world communication language language gets to say and who gets to decide. Perspectives and power. Rules
by others and roles.

Proposition Three based questions

Direct observation of Indirect data we get What information is available to those who have an interest in

the physical world from others

the activity, and where is it sourced?
What information is necessary to inform the activity that will
ful l those needs?
Where are people getting their information from? If it is ltered
Information we use through other people, is it being distorted in this process? If so,

to decide what to do
why and how? Do they have any veri cation strategies for
indirect information?
What values and assumptions are underpinning the ways in
which people are processing and analysing information?
Language is where many communications problems lie. One person Is information from some sources given more weight than
may have perceived what is going on incorrectly, or have interpreted others simply because of the power and status of the source?
what she saw in very personal ways, but she may also
misunderstand the words that a person uses to describe what they
experienced.

In any focused group activity all four sources types of information are

System diagrams 89
fi
ff
fi
fi
fi
Example – Horticultural Activity System Humans are tool users. A tool is anything you use to help you
manipulate the environment to meet your needs. We also use them
The cancellation of the contract with the major supermarket chain to obtain needed information from the environment.
came as a complete surprise to Local Stu . The chain had been
honest in providing information indicating they were considering Physical tools extend our physical capabilities and conceptual tools
alternative suppliers, but Local Stu had assessed that information extend our mental capabilities. Both have embedded in their design
based on the assumption that the contract was secure because of the ideas and assumptions of the people who developed them.
their long relationship with that supplier. In an example of the data These always have limitations. For example, the desk I am sitting at
not speaking for themselves, the two parties were interpreting the as I type this is not mine. It has a lever to adjust its height. However,
same information through di erent value systems and assumptions. the position of the lever assumes that the desk user is right handed.
But I am left-handed. Every project or program has embedded in its
design assumptions that the developers have about the users’ world.
For example, when developing this book I sent out a draft for
Proposition Four – Tools comment that assumed people would use a personal example
alongside my text. I assumed they would treat is as a workbook not a
textbook. However, some people responded as if it were a textbook
We use tools to manipulate our environment and to get that you read independently of any speci c example. Consequently,
information from the environment. The tools we use mediate (or that feedback did not entirely ful l my needs.
shape) the way we do the work.
Indeed, language and writing are perhaps the most important
conceptual and physical tools that humans use. And they too have
their limitations as the above example highlighted.

Tools and
So while we use tools to manipulate the environment, the nature of
the tools we use also shapes our own thinking about what to do and
Artifacts how to do it

mediate Proposition Four based questions

People Environment Needs

What physical and mental tools are needed for the work? What
knowledge and skills are needed? Are they present?
acting on to satisfy Are the tools in use well suited to ful lling the stated needs?
In what ways are the tools constraining or in uencing the work
geared towards ful lling those needs?

System diagrams 90
fi
ff
fi
ff
fi
fi
ff
fl
 Do stakeholders have su cient skills to use the available tools Rules – the formal and informal principles and directives that set a
e ectively? Do they have the relevant literacy and language boundary around our actions directed at ful lling needs.
skills?
Community – those to whom we most closely identify with, live with,
Are some stakeholders privileged over others in the use of the work with and generally engage with in activities focused on ful lling
tools? Does this matter, or is it merely a sensible division of needs.
labour?
Organisation – the roles, structures, hierarchies and processes
within which activities take place.
Example – Horticultural Activity System
Tools and
Artifacts
Clearly this activity system uses a lot of tools. Perhaps the most
important are ‘quality control’ tools, and ‘relationship’ tools such as mediate
legal contracts. An example of issues managing the complexity of People Environment Needs
these tools was the language used in the contractual relationships acting on to satisfy
over quality control tools. This was the major reason for the
supermarket contract cancellation. Maybe Local Stu compared with Rules Community Organisation
the supermarket chain had inadequate skills to develop more robust Formal Those around Work practices/roles
contractual relationships. I nformal us
(cul ture)

Proposition Five – Collective thought and action

The human systems – cultural, social and organisational – also The above diagram represents these additional dynamics operating
mediate the ways in which we conduct our activities. The cultural within an activity system.
is mediated via rules, the social by communities of practice, and
In summary, people act on their environment to ful l a need. The
the organisational by the division of labour or roles.
nature of that work is mediated by the rules they are operating
under, the traditions, relationships and expectations of their
It is not only the tools we use that shape how we approach our work.
community and the various roles that they have been allocated or
We also ful l our needs through social practices. CHAT identi es
acquire.
three key social practices, rules, communities of practice and
organisational roles that help us meet our individual and collective Therefore the rst step of any activity system based inquiry is to
needs. identify the shared focus of the work.

System diagrams 91
ff
fi
fi
ffi
fi
fi
ff
fi
fi
 work? What needs are they seeking to address?
It is a major aspect of CHAT that people will always have di erent
What are the formal rules; manuals, standard operating
reasons, di erent motivations, di erent needs for engaging in a
procedures, etc., that promote or constrain the way in which
shared activity. Indeed, as you will see later, these di erences can be
the activity proceeds?
a major source of di culty or learning, depending on how they are
managed. What are the informal, cultural, rules that promote or constrain
the way in which the work proceeds?
For instance, let’s assume we are working on a needle exchange What are the organisational structures that shape the way the
scheme, where drug users can exchange a dirty needle for a clean work is done?
one. Your purpose for working there is because many of your friends
have died being infected with HIV from using dirty needles. You What other systems must supply inputs in order for the work to
need is to address your grief. My purpose of working there is to see proceed? What other systems use the product of our systems’
if I can score some drugs in a relative safe environment. My need is work? How are all these systems connected to our system?
to satisfy an addiction. One of our colleagues is there because she What information must ow around the system for it to operate
is just out of college and needs something interesting on her CV to e ectively? Where is the information needed kept, and where
attract potential employers in the public health eld. She needs must it go? Is the information owing as required?
employment. The outcome of our collective work is the reduction of
a collective need, in this case blood born diseases. However, we are
engaging in a collective activity – operating on the environment – to Example – Horticultural Activity System
ful l di erent individual needs. The next two Propositions will
explore this in more detail.
The Local Stu activity system has a rich mixture of role, rules, and
communities of practice.

Proposition Five based questions Roles (organisational and individual)

What is the need being satis ed by the outcome of the Consultancy

activity? What goods or services are being produced through Retailer
the activity to ful l the need? Food producer
Driver
What raw materials and prerequisite conditions are required Supermarket shareholders
for the activity to start from? What are the processes by which
the raw materials and the prerequisite conditions are
transformed by the activity to serve the needs?
What are the di erent kinds of people necessary to do the

System diagrams 92
ff
fi
ff
ff
ff
ff
fi
ffi
fl
fi
fl
ff
fi
ff
ff
 Rules systems approaches and mentioned several times in this book; the
role of dialectic. In CHAT parlance dialectics are described as
The system must grow.
disturbances or contradictions.
The businesses involved must be pro table
Local Stu should have a stable workforce. When the tools, rules, community and organisation operate as
expected, those within an activity system proceed by conducting
Producers get a fair price for their produce. standardised tasks with predictable results. But the system will often
Retailers pay a fair price for the produce. be interrupted by unanticipated events (disturbances), or surface
underlying tensions between elements of the system
Good and constructive relationships between producers, (contradictions). When a team, program or organisation encounters
consultancy clients, retailers and Local Stu . these it will descend into chaos unless it is able to learn how to deal
Local Stu is worker-friendly and a good place to work for and with these issues.
to work with.
Retaining high quality produce through the process chain is There are four possible sources of disturbances or contradictions:
essential. Within components of an activity system (e.g. between rules)
Between components of an activity system (e.g. between rules
Communities of practice and roles)

Sustainable farming Between activity systems. A performance measure used to

ful l an organisation’s need to reduce injuries, may contradict a
Sta management performance measure used to support the same organisation’s
Business development need to sell product. Indeed contradictions in performance
measures are common systemic contradictions.
Historical disturbance – between what is now and how it used
to be. For instance, between a newly introduced tool and an
Proposition Six – Learning old rule or role. The introduction of desktop computers (tool)
in the 1980s ended the role of typist and the typing pool (role
In CHAT disturbances and contradictions in system components and division of labour).
are considered opportunities to allow you to learn about a
situation rather than just problem solve.

As mentioned in the introduction to this chapter, CHAT is based on

theories of learning. Here I highlight one that is common in other

System diagrams 93
fi
ff
ff
ff
fi
ff
 In my own practice I nd it useful to identify and explain (rather than individual to organisational to societal – that are not always explicit.
explain away) four kinds of contradiction:
Such analyses help you to understand why people do things the way
they do and why those observable actions often seem to be in
An exception to a generalisation that can be expressed as ‘in con ict with the stated desired results of the activity system.
general … but …’
Once the shared need(s) has been identi ed, the next step is to
A true contradiction that can be expressed as ‘one the one analyse the structure of the activity system using a number of
hand …, on the other hand… guiding questions.
A surprise that can be expressed in two ways: ‘I expected to
see X but it wasn’t there’, or ‘I didn’t expect Y to happen but it
Proposition Six based questions
did’
A puzzle, expressed by ‘I really have no idea what’s going on’.
What kind of contradictions are there within the system? What
The most di cult contradictions to use as springboards for growth have been the consequences? How have people responded?
are those that are invisible or undiscussible. How could they have responded?
An invisible contradiction is one that is so much part of a team’s What generalisations do people make about the performance
everyday life, so taken for granted as a given that members don’t of the system? What exceptions to those generalisations are
even recognise it as a contradiction. Invisible contradictions include there? What learnings are there from these contradictions?
anything that is taken for granted. Cultural assumptions about how
things are done and how relationships are managed are examples. What contradictions are there? Where can you state ‘on the
one hand and on the other’? What were the consequences?
Undiscussible contradictions are those that nobody ever talks about What are the potential learnings?
because they are embarrassing, uncomfortable or culturally di cult
to confront. Gender and racial issues in teams, or o ensive personal What disturbances – unanticipated events have happened?
habits of politically powerful program stakeholders, are all examples What were the consequences? What are the potential
of undiscussible contradictions. Nobody is willing to talk about them learnings?
openly, but they may be seriously impeding progress towards a What are the historical underpinnings of these contradictions
result that will ful l the identi ed need. and disturbances? How is the past interacting with the present,
Surfacing invisible or undiscussible contradictions, and stimulating a the old with the new?
developmental dialogue around them is the most potentially What events and circumstances remain undiscussible?
valuable service that a CHAT based intervention can provide. Between whom are they (un)discussible? What rules, roles,
Analysing these multiple objects, and surfacing the tacit or unspoken tools, objects and histories mediate these undiscussibles?
ones, is an essential task. But at the tacit level tensions between
objects almost always produce multiple levels of tension – from

System diagrams 94
fl
ffi
fi
fi
fi
fi
ff
ffi
 Example – Horticultural Activity System has strong direct connections within the drug using community.
Over the years the activity system has been confronted by massive An airline will seek to transport passengers safely and reliably, but it
and simultaneous sets of contradictions. There were contradictions is often the case that operational trade-o s may need to be made
within system elements (especially between rules, and between the between reliability (e.g., arriving on time) and safety. The calculations
desired results of the activity system ) and also between system are rmly embedded in routine airline practice. But then suddenly
elements, such as between roles and tools, between rules and tools, there are hurricane warnings over hundreds of square kilometres.
as well as some historical contradictions including historical bad Flights need to be diverted or climb away from the storms in order to
relationships between Local Stu , retailers and producers. reach the destination at all.
The rst major disturbance was between the growth rule and the CHAT is commonly used to explore health and safety issues,
rules baked into the Local Stu structure. The Local Stu because there are so many unspoken, tacit, reasons why a group of
organisation became too big to sustain the historical way in which people are undertaking the same activity, and the legal,
produce was distributed. The new e ciencies within the system led organisational, group, social and cultural contexts within the activity
to the ability to supply large quantities to a single retailer. system. Some help, some hinder safe and healthy results. It is well
known that unaddressed contradictions between di erent
The next disturbance occurred as a consequential contradiction organisational needs, rules, roles and tools are major causes of
between the growth rule, the information tool and some aspects of workplace injuries.
the overall purpose of the system. Responding to some
contradictions of its own related to quality and loyalty, the major Many years ago I was involved in an evaluation of a workplace injury
retailer dumped Local Stu at short notice, something that the prevention program. In one case study, we came across a really
systems information tools were unable to pick up, possibly because caring and competently run construction company that prided itself
of an over-reliance on the good relationship rule. on two things. The well-being of its workers and delivering jobs on
time. Neither of these are exactly common in the building industry.
And it also resolved a tool contradiction by changing its role. Local Consequently it was devastated by the serious injury of one of its
Stu survived by expanding the activity system and its actors by workers. When we unpicked what had happened, it became obvious
introducing a highly pro table consultancy business into the system. that the worker was seeking to resolve a contradiction between
In other words it addressed a tool problem not by changing its tools, these two organisational values (rules), and traded-o a common
but rather by changing a role. safety practice so that the job could nish on time. It is a classic
example of contradictions between high level rules can a ect
ground level practice.
More examples
In the needle exchange example, a disturbance might be the
introduction of a rule that existing drug users can no longer work in
the project. If that is not resolved I may leave the project, which
would create a further disturbance, because I am the only one who

System diagrams 95
ff
fi
fi
fi
ff
ff
ff
ffi
fi
ff
ff
ff
ff
ff
 Proposition Seven – Development demonstrates how learning through addressing cascading
contradictions can be expansive.
The consequences of action may lead to further contradictions
It is not only how to learn from contradictions or disturbances that
within the system, such as a new tool imposed on old rules. It could
matter. At the heart of CHAT is the principle that a contradiction is
even go beyond the boundary of the existing system and explore
an opportunity for participants to enter a Cycle of Expansive
contradictions between adjacent activity systems. The earlier
Learning.
example might be a health and safety activity system and a business
pro tability activity system. In doing so this expands the possible
boundary of learning beyond the initial focus of the inquiry.
Another diagram:

Fourth-order contradictions
(between activity systems) Proposition Seven based questions
First-order contradictions
1. Present practice: (within the elements What is the history of how current activities came to be as they
In-need state of the activity system)
are now?
Third-order
contradictions What kinds of weaknesses exist in the relationships between
Second-order
(between contradictions the elements of the system?
the old 4. Application and (between the
and the generalization: 2. Double bind: lements of What is missing that is needed?
new way) Changing Analysis and search the activity
the activity for a new solution system) What is not working as well as it might? What relationships are
system
not working as they should?
What strengths are there in the system? Are they being used
as well as they could be?
What potential for growth and development is there in the
system? What is desirable? What weaknesses and de ciencies
need to be recti ed before the potential can be tapped?
How could possible changes impact on the existing activity
system? What are the learning opportunities and how can they
The above diagram displays how disturbances provide the be enhanced?
opportunity for learning and innovation. If disturbances are seen as Are adjacent activity systems likely to be a ected? If so, in
irritations, problems to be overcome, then it is unlikely that what way, and how can these learnings be exploited?
stakeholders will progress beyond re ghting. The diagram

System diagrams 96
fi
fi
fi
fi
ff
fi
 Example – Horticultural Activity System
Local Stu sought to resolve a role disturbance by reorganising its
CHAT resources
role in the system from broker to becoming the storer and deliverer
of produce. Local Stu responded essentially by changing the This chapter is substantially based on an original paper Activity
community of practice, it altered the relationship between the rules Theory As A Design Principle For Team Development Processes by
and roles by essentially changing the division of labour. It laid o Capper, Hill, Wilson, Harvey.
sta and changed the way in which food was stored and delivered. It is available here:
In that way it expanded the community of practice, but kept the good
http://www.bobwilliams.co.nz/ewExternalFiles/CHAT.pdf
employer rules by altering whom it employed. Thus the solution
allowed the rules, tools and certain actions to remain largely A key source is the Center for Research on Activity. Development
untouched. and Learning at the University of Helsinki
https://www2.helsinki. /en/researchgroups/centre-for-research-on-
However, as with any cycle of expansive learning, this resolution
activity-development-and-learning
created further disturbances. The new division of labour created
tensions within rules and tools around the need to maintaining high This is also an excellent primer:
quality. A new set of tools had to be developed, which also allowed http://faculty.washington.edu/kfoot/Publications/Foot-CHAT-explored-
the quality control rule to be expanded and is probably now best dist-tf.pdf
expressed as ‘consistent quality in a changing environment’.

System diagrams 97
ff
ff
ff
fi
ff
 So Where Do You
You might think that a rather gloomy prognosis for how you can
respond to the ideas in this book.
But it doesn’t have to be framed that way.

Go From Here?
There is an alternative.

Think systemically, be systemic

Why not frame the application of systems approaches as itself a task
that requires some systemic thinking?
And because system diagrams are at the less challenging end of
Some advice from the trenches thinking systemically, they provide a useful rst step to being
systemic.
System diagrams, systems thinking and systemic approaches to Take a look at this Rich Picture from the Open University system
problem analysis and resolution are increasingly popular. The diagram course material:
systems eld has been in this position several times over the past
century. It is a predictable cycle. There are a bunch of problems that
seem insurmountable, someone recollects learning something about
systems somewhere, there’s a rush for adoption, expectations run
high, disappointments inevitable. Systems thinking fades into last
year’s fad.
This is no surprise. The reality is that systems thinking poses big
challenges to the way that people see the world and address real-
world problems. Well established structures, processes, power
relationships, and practices are called into question. The forces that
drive the adoption of systems ideas can be overwhelmed by the
changes necessary to do it.

System diagrams 98
fi
fi
 Or why not consider drawing an In uence Diagram of what might Six Simple Rules for those First Steps
in uence the adoption of systems ideas into your life.
Or on the other hand, why not use a checklist that I have developed
over the years in seeking to apply systems ideas to new situations:
Maybe you are con dent enough to consider a Causal Diagram that
could identify possible leverage points?
Use an approach that interests you. Enjoy yourself.
Find someone in that situation who is also interested and has
Or maybe the Cyne n diagram can get people to think about how
some authority.
your assessment of the situation lends itself to di erent ideas and
how they might be introduced. Start somewhere safe. Seek low risk, medium reward before
high risk, high reward.
Reassure yourself and others that the approach you chose
In VSM terms, can you identify where a diagramming approach may
really could make a di erence or a unique contribution.
help to reveal or address variety imbalances in your situation?
Do your homework, make sure you know what you are doing.
Adapt, modify but make sure that you don’t stray too far from
Perhaps CHAT may highlight what rules or roles in which the basic principles of the approach you choose.
communities might need to be present to use the systemic thinking
tool on a collective task. Or consider what are the undiscussible
contradictions with using systemic approaches and how they can be The key thing is to give it a go. Using diagrams is a hands-on task.
made discussible? You can understand the basics by reading books like these, but you
develop your skill by doing.
So the sixth rule is to pick something, just do it.

And have some fun doing it.

System diagrams 99
fl
fi
fi
ff
fl
ff