Description
Probabilistic forecasts are used for everything, from predicting tomorrow’s weather to generating betting odds on sporting events. Instead of looking at one possible outcome, this technique looks at all possible events and assigns a probability of each one occurring.
The key insight here is that instead of pretending to know exactly what is going to happen in the future, a user acknowledges they are not omnipotent, and that the future is uncertain. This approach can be particularly useful when forecasting for our supply chains, as there are multiple uncertainties and many things that we simply don’t know. By taking a probabilistic approach, we can capture some of this “fuzziness” and allow for a sharper, more logical reasoning of future events.
But what does this actually mean in practice and how does it compare to more classic techniques? When it comes to more traditional forecasting, you usually make a single statement for the future - a prediction -, then wait and see how far your result differs from what actually happens in reality. This difference between the real and virtual world is referred to as variance and is often what management ends up getting somewhat irritated about when they find that too much or too little stock has been bought.
The puzzling thing with these classic approaches that look at only one single future is that they don’t at all account for real world uncertainty. This is where a probabilistic approach can help, as it’s the extremes that occur in real life, creating the dips and spikes on your graphs that need to be looked at in more detail, because this is where over-stock and under-stock scenarios actually happen.
In conclusion, although probabilistic forecasts may sound highly technical and intimidating, many supply chain practitioners have usually already been doing a similar sort of prediction for years. By relying on their experience, knowledge of their stock and “gut feeling”, they organise the likely from the less likely possible scenarios. This experience and innate knowledge can then be enhanced by technology to further refine forecasting and have a better managed supply chain.
Find out more about Probabilistic Forecasting
