Great theories come in two sorts: the type that is so fundamentally foreign to one’s existing worldview that it’s a major effort to get your head round it (think Special Relativity here); and the type that is so staringly obvious that you can’t believe you didn’t think of it yourself – but you didn’t. Both types enhance your perspectives and help you to see things afresh, discover stuff you’d not otherwise have noticed and/or provide you a theoretical framework to justify your own intuitive feelings.
W. Ross Ashby’s Law Of Requisite Variety is one of the second kind. You don’t need a degree in physics (or anything else for that matter) to understand it. There is in fact some serious maths associated with it but you can ignore that (for now). It just makes sense! So what is this Law then? If we look at any system (an organization, an organism, a mechanical or digital system), we can see that it’s constantly confronted by events, which can alter the behaviour of the system. These events can be internal or external to the system. The total of these types of event (or factors) is the Variety affecting the system. What Ashby says is that, if you want your system to remain viable, you need to match that variety with variety in your ability to monitor and respond. In other words, for every factor outside of your own control that can have a positive or negative influence on your ability to meet your goals, you need to put a compensatory control in place to allow you to respond.
It seems that very few people outside of the world of cybernetics were aware of it or him until the last few years. I certainly wasn’t. The somewhat better known Stafford Beer did apply it to organization theory (management cybernetics) with his Viable Systems Model but Beer’s work too suffered years of neglect after the CIA and their friends (literally) killed off his experiment with the Chilean economy in 1973. The current wave of interest has taken place in the very practical worlds of management science, systems thinking, social business and enterprise architecture. Patrick Hoverstad in his The Fractal Organization uses Requisite Variety consistently to explain why simplistic, top down management models have failed time and again. Dave Gray uses it in The Connected Company. Ruth Malan has named her wonderful architecture discussion page after it (with good reason) and Larry Constantine wrote a short story about it. The list goes on.
As it happens, the simplicity of the Requisite Variety concept as stated here (or for that matter the more formal looking definition in Wikipedia), allows a bit too much interpretation and in particular over-simplification, which undermines the importance and value of the law. One symptom is trying to make the law into a set of rules.
Ashby himself pointed out that you don’t have to (and can’t always) know how the influencing factor will behave. Only why it might matter to you, so you’re able to respond when something happens. That in turn means that your responses can’t be predetermined. Or to put it another way, the law is not deterministic. Ruth Malan took this a step further recently and introduced the idea of “requisite flexibility“, meaning that not just that you respond but how you respond needs to be able to adapt to changing circumstances or changing goals.
Another symptom is stating the law but failing to explain how to use it. It’s all very well recognizing the truth of the statement. The question is how we are to act on this truth to create a framework for viability in our organizations. How are we to identify all that variety, understand its potential impact and create mechanisms to allow us to respond? If we can’t do this, we’re not really much further than knowing that falling stones hurt if they land on your head.
I’ve been reading through the available literature trying to answer three questions.
- What actually constitutes variety;
- How do we go about discovering it and
- How can we determine what the right response mechanisms are?
In the next few blogs I’m going to look at various approaches that have been developed to address some or all of these questions. I hope in the end to have found enough to cover all three. This isn’t going to deliver a deterministic answer. In fact it isn’t going to deliver an “answer” at all. Like all good theories and insights, Ashby’s law helps us construct our own methods and/or frameworks that can be applied to different kinds of concrete situation and deliver results appropriate to that situation. Remember Ruth Malan’s requisite flexibility principle.
As Tom Graves has observed elsewhere , the law is in a sense recursive, as we need to apply it to our own use of it too. If our analysis methods don’t allow for variety, we’ll commit the classic error of trying to force fit reality to a rigid method. So guided by Ruth and Tom’s ideas, you might say that I’m about to embark on a quest to save Ashby… from his friends.