There is a lot of talk in the business media about uncertainty and however well-meaning, there seems to be a general message that we just have to accept uncertainty, which to a degree is true, but such a viewpoint can also become an excuse for costly inaction . It turns out that ‘uncertainty’ is a more complex beast than we are led to believe, and that this is a good thing because by understanding uncertainty we find that we can in fact do a lot better at reducing it then we may think.
What is Uncertainty?
Uncertainty is a subjective idea that represents a general lack of ability to know what is going to happen at some point in the future. At any point in time a complex system has a variety of possible sates or outcomes that it could move to and as its complexity increases the number of such states and the rate at which they increase can rise dramatically. It is this pool of potential outcomes that we see as the, ‘intrinsic uncertainty’ of a system. Simple systems have less uncertainty because they have fewer options or less variety in their behaviour.
It is this, ‘intrinsic uncertainty’ that commentators talk about when they suggest we should, ‘embrace uncertainty’ and to a degree they are correct. However what tends to get ignored is the fact that there are other sources of ‘systematic uncertainty’ that we introduce and this combines with the intrinsic uncertainty to increase our overall uncertainty in a given situation. And so in a business context; things tend to be more uncertain than they need to be.
The truth is that we can do a great deal to reduce the influence of the systematic uncertainties in our businesses, but more surprisingly we can also take action to potentially reduce the intrinsic uncertainty, however we can never be less uncertain about the outcome of an event than the intrinsic uncertainty and everybody has to accept this because that’s how it is. I call this, ‘the business uncertainty principle’ and although we never get anywhere near this level of uncertainty in the real world we can apply approaches to dramatically reduce the uncertainty that we have become accustomed to.
Reducing Intrinsic Uncertainty
There are two approaches we can take to reducing the intrinsic uncertainty that is generated by the complex behaviour of our businesses. The first strategy we should adopt is to have a relentless pursuit of simplification and take the view that complexity always has to be justified and not seen as a consequence. Now it is true that simple systems can exhibit highly complex behaviour, but in general simplifying your business operation is likely to reduce the complex behaviour in certain circumstances and at certain times. There are many aspects to simplification, but in general the approach should be to reduce the influence between elements and to reduce the number of elements within the operation. A particular example of this influence is positive feedback, where the influence of one activity is somehow fed back into the activity to influence its future behaviour. This loop of influence can have an amplifying effect causing the behaviour of the activity to increase in complexity.
The second action we can take is based upon the idea of regulation which originated from the study of cybernetics. Regulation in this context means maintaining the behaviour of the system within bounds so that the more rapid and larger magnitude swings in behaviour are reduced through rapid dynamic intervention. The most pragmatic and effective way to do this in a business operation is to optimise localised adaptation and this means giving individuals the support, training, responsibility and authority to respond to whatever is going on around them. This approach has some support from complexity science in that localised adaptation can reduce the periods of higher levels of complex behaviour and hence reduce intrinsic uncertainty.
One thing we have to be aware of is that large scale change that is focused on adapting the behaviour of the business has by definition more uncertainty in its outcome because it takes the business a certain time for the influence to be propagated and in this time it may have naturally moved away from the state for which we applied the change. In essence with large scale changes we are continually chasing our tails.
Reducing Systematic Uncertainty
We can now address the sources of systematic uncertainty that all combine with the intrinsic uncertainty to produce what we see as a compound uncertainty.
The idea of uncertainty is related to our ability to make predictions of future behaviour within acceptable levels of accuracy, so the less effective we are at prediction then the more systematic uncertainty we are introducing and it has always amazed me how little attention we generally pay to prediction. The primary way to increase our ability to effectively predict a system’s behaviour is to model the dynamics of the specific aspects that we wish to be more certain about. To do so we need to become more ‘system aware’, as opposed to focusing on the characteristics of an individual aspect of the business operation.
The amount of potential variation (intrinsic uncertainty) that a system has and the rate at which it changes its state increases as the complex behaviour increases. This means that the rate at which our predictions diverge from reality rapidly increases as its behaviour becomes more complex. This means that our predictions have a period in which they produce acceptable results that I call the ‘prediction window’. My own experience is that in general we do not predict or estimate often enough and that our models need to be dynamic and re-assessed with the latest real data as regularly as possible and this requires us to move toward computer based simulations to help us improve the effectiveness of our predictions. These predictive models will be complicated because we are trying to model a complex system, however they are still easily manageable and I have personally developed specific models using just Microsoft Excel that have greatly improved my ability to predict likely behaviour for given scenarios, and with some tinkering such models can be re-used on other projects. To achieve this improvement in prediction we need many more managers who are system thinkers with modelling skills and who have an understanding of the basics of complexity science.
If we get to a point where we have standard and regularly updated prediction models we then we can become more sophisticated in that if we measure the period of time in which our predictions remain acceptably accurate then we can start to monitor whether our system is becoming more of less complex in its behaviour.
Another source of systematic uncertainty is our limited ability to make objective judgements and decisions. Every decision we make will have a propagating influence upon the business and can help drive complex behaviour. When we make decisions we tend to focus on the anticipated result and not on the wider influences that such decisions can have. Once again our models can help us develop a better understanding of the consequences of actions and, ‘what if’ analysis should be part of everyone’s toolbox. In general higher level decisions will take longer to propagate through the business and are likely to increase the intrinsic uncertainty in given scenarios.
Studies have suggested that even using basic approaches to objective decision making such as using decision trees or even basic scoring methods greatly improve our judgements. From a complexity science perspective one of the options that we often neglect is to do nothing and accept a period of hurt and let the system self-regulate and modelling can help us see if this is an attractive option.
Be Aware of the Complexity Storm
Finally there is one other aspect of uncertainty that applies specifically to project based operations that is useful to be aware of. Projects consist of an escalating number of interdependent tasks. This increase in active tasks tends to accord with an increase in the complex behaviour of the project. For some period during the project the complexity reaches a peak that I call, ‘the complexity storm’. During this period the intrinsic uncertainty is at a maximum and we need to be aware that this is the time of maximum uncertainty within the project. Unfortunately the project’s sensitivity to change increases with its complexity and so the actions we take during the storm have a greater likelihood of creating even more instability. My own experience tends to suggest that if a project is going to fail then it’s during this storm. From some of the measures mentioned above coupled with a general understanding of the work breakdown structure of a project one can identify when the complexity storm is likely to hit and this can help to moderate one’s expectations and positively influence the decision making process.
One of the reasons why Agile based projects are successful is that they retain a level of relative simplicity and a task structure that smears out the complexity storm to create a much less intense but continual complexity shower that is more stable to change.
How to get Organised
So given all this, how does an organisation move forward? Well it may be that organisations need to create a new business function that has a responsibility to promote training, techniques, methods and tools to combat uncertainty in all its forms across the organisation.