The VSM Guide

An introduction to the
Viable System Model
as a diagnostic & design tool
for co-operatives & federations

Jon Walker

Version 3.2 (2020)

Section 0: Cybernetic Eyes
Section 1: The Quick Guide to the VSM
Section 2: Case Studies
        Hebden Water Milling 1985
        Triangle Wholefoods 1986
        One Mondragon Co-operative 1991
Section 3: Preliminary Diagnosis
Janus interlude
Section 4: Designing Autonomy
Section 5: The Internal Balance
Section 6: Information Systems
Section 7: Balance with the Environment
Section 8: Policy Systems
Section 9: The Whole System
Section 10: Application to Federations

Appendix 1: Levels of Recursion
Appendix 2: Variety

Creative Commons Licence
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.


Janus has rightly been adopted as the god of whole systems. He can look forwards and backwards simultaneously, and thus symbolises the closed loop at the heart of all cybernetic systems.

This Janus interlude is included at roughly the halfway point in the book to provide the reader with a perspective on where he's been and where he's going.

Looking back ...

Having completed the Preliminary Diagnosis, you should by now have a thorough feel for the form of the model.

  • The separation of Operation, Metasystem and Environment

  • The small VSMs nested recursively within the Operation

  • The four Metasystemic functions

If you have been drawing the pictures, you should by now have a number of large VSM sketches, showing the five systems and the bits of your organisation which do those jobs, and you will probably have had a Eureka! or two, as the VSM version of organisational reality unfolded.

Looking forward ...

So where do we go from here?

I admit that my first few attempts at using the VSM finished here, despite a few nagging doubts about Beer's Laws and Axioms, which didn't really seem to fit in anywhere.

After much further reading and discussion it became clear that the job had only been half-completed. YES, I had identified the five systems at various levels of recursion, but NO, the diagnosis was by no means complete.

Consider now, the Viable Vehicle Model, or VVM, a guaranteed fool-proof methodology for the diagnosis and design of any vehicle, regardless of size or the terrain on which it must travel, based cybernetically on the invariances in vehicular functioning.

The VVM is composed of 5 systems:

  1. Some means of propulsion
  2. Some means of steering
  3. Something to carry the load
  4. A method of reducing friction
  5. Some way of stopping

Imagine you have completed the Preliminary Diagnosis of your Vehicle-in-Focus (or VIF) that you have successfully identified the five systems, and drawn a big diagram to illustrate their basic relationships.

So what now? Is the VVM diagnosis complete? Again you may have a few nagging doubts about the VIF as a whole. For example, you may be concerned that your design of your Propulsion System One (in this case a very substantial rubber band) is not quite up to the job of moving the Vehicle with the 400 tonne loads you envisage. Similarly System 5 (a large rock on the end of a rope) has been clearly identified as a fine example of the stopping function, but as the vehicle is intended to transport steel in Alaska (and thus has System 4 consisting of skis) doesn't look too reliable.

The overriding questions begin to emerge: Do the parts add up to a workable whole? Are the parts balanced to complement one another?

VVM practitioners will know that the workings of the whole is a very different thing from the workings of the parts. In some cases overall improvements in efficiency may be the result of actually decreasing the capabilities of one sub-unit.

So, if the vehicle is on wheels, are the brakes and engine designed to match the weight and speeds for which the vehicle is designed? If the vehicle is a hovercraft, the engine could be smaller as friction is minimal, but this may present more problems for the design of the stopping and steering systems.

The conclusion is:

  • The design of the whole system must be considered.

  • Just because the parts are there, there is no reason why they should work together to produce an effective whole.

And exactly the same is true of the VSM.

In my initial studies, I had not looked at the way the parts interact, how they work together to generate a well-designed whole, how the capabilities of the regulatory systems balance the problems they are dealing with.

Everything which follows is about balance.

The essence is well-balanced, whole-system design.

This is where the forward-looking head of Janus is looking.