Cybernetics assumes that this basic architecture underpins every system that behaves in a purposeful way. I will start by describing how this works for a very simple mechanical system – a domestic central heating system - and use this to illustrate how the same mechanisms manifest themselves in an organisational context, with all the attendant complexity.

 

For convenience let’s start with the environment that perturbs the system, which is in this case the external temperature and its impact on temperature within the house. Sensors detect the internal temperature and send this information to the regulator i.e., the thermostat attached to the boiler. The thermostat compares the actual temperature to the desired temperature and sends a signal to a boiler to fire up if the temperature drops below the goal.

 

The fundamental law which governs the operation of systems like this is Ashby’s Law (learn more here). This tells us that this system will work well if the range of regulatory responses and the range of desired states is in balance with the range of variability in the environment. If the goal states are too tightly defined and/or the range of responses from the regulator does not match the rate at which the environment changes the system will have failed…i.e., the house will be either too hot or too cold.

 

Let’s now start adding in some complexity. 

 

Imagine for the moment that the regulator cannot heat up the house quickly enough for its occupants. One way to compensate for this is to attempt to anticipate changes in the weather so that the boiler can be fired up before the sensor records an actual temperature drop. In effect we have created a forecast using a weather model to create ‘future actuals’ to drive the system.

 

Even though this is still a stupidly simple example, all the features of an organisational control system can be clearly distinguished. 

 

The desired temperature settings we call ‘targets’, the regulator is the decision-making process that drives the allocation of resources which is informed by the sensor that reports on actual performance and the weather forecast that anticipates the future house temperatures.

 

And it is also clear why Beyond Budgeting succeeds where traditional approaches, based on annual budgets and single point targets fail. 

 

Put simply, goals are more broadly defined, and the regulatory system is more responsive to changes in the external environment making it more likely that the system will deliver the performance required. This is what is described in Little Book of Beyond Budgeting and my other books address the issue of evaluation performance and goal setting (Present Sense), forecasting (Future Ready) and resource allocation (Cost Matters).

 

At this point however the analogy breaks down because the task of leaders and managers in an organisation is impossibly more complex than that that any mechanical system can cope with. Let’s uncover some of the sources of complexity.

 

Evaluating Performance.

 

In our central heating system, we are attempting to measure one variable, as continuously measured by a single sensor in one room in the house. 

 

There is an almost infinite number of things that are important for organisation health. Many of them can’t be measured at all, and those that can be measured are not or cannot be tracked continuously and they interact with each other in ways that cannot be easily untangled. 

 

Also, unlike temperature which can be measured to a very high tolerance most organisational data in infected by an unknown amount of noise, so it is very difficult to establish the true signal.

 

Setting Goals

 

For a central heating system, it is relatively easy to set a goal since the thing that determines what is acceptable is our body temperature which is constant, and we have other strategies that we can use to compensate for failing in the heating system, such as putting on extra clothing. Also, relatively minor deviations are not fatal, they are just uncomfortable.

 

On the other hand, there are many things that an organisation might want to target and many ways in which we might want to target them. Some we might define in broad ranges, others might be ‘fatal’ if we breach certain well-defined limits, for example having no cash.

 

Also, the relative importance of different variables and target level will change over time. For example, ‘good’ may be defined by performance relative to peers rather than is absolute terms. So, market share is a better measure than sales and for publicly quoted companies poor relative share performance will increase the risk of predation

 

Forecasting

 

I know of no central heating systems that is driven by forecasts, but it is difficult to find an organisation that does not rely upon them to some degree.

 

Forecast have all the problems and complexity involved with evaluating performance and some more of its own. 

 

Every forecast requires a model of the organisation in order to estimate how a change in an input will be reflected in the output we are interested in i.e. performance. The challenge is that organisations are very complex which makes it difficult to forecast output. Also, there are often many different latencies; some things will have an impact very quickly, others will take much longer. Worse still the nature of the organisation is not fixed, so we can’t guarantee that past behaviour will be reflected in the future. 

 

Finally, we also need to forecast the input which is usually be very difficult. What will happen to the economy, our markets and how will our competitors react? 

 

Allocating Resources

 

If the actual or forecast performance of the organisation is not in line with our aspirations, we will need to change something, and this usually means deciding to allocate resources differently.

 

This process has the same challenge as forecasting. We need to know how a change in the input (resources) will impact performance (the output) and when which because the organisation is so complex and changing is very difficult to predict.

 

Scale

 

All these challenges are faced by a small one-person business. But for bigger businesses the level of complexity increases exponentially. 

 

There are thousands of control loops, some more important than others, constantly interacting with each other in ways that are difficult to predict. As a result, cyberneticians tend to treat organisational entities as ‘black boxes’ simply seeking to understand the relationship between inputs and outputs rather than explicitly modelling the causal structure.

 

In addition, organisations need to do more than respond to perturbations in the existing environment in the short term, they need to adapt to fundamental shifts in the environment in the future; a process that can also be modelled as a causal loop.

​

People

​

The most important source of complexity in larger organisation is people. In fact, social organisations are people.

 

Every individual has their own economics and psychological needs and aspirations which they will pursue through work. This can be a source of opportunity if these are aligned with those of the organisation and a major problem when they are not.

 

In addition, social controls (what is commonly called culture) operate in the same way as the explicit procedural control loops we have discussed thus far and like them are subject to the contracts of Ashby’s Law.

​

The people dimension is touched on in all my books but the interdependent nature fo social and procedural controls features most prominently in 'The Viable Model Workbook'.

​

​

Discover more about cybernetics, starting here.