Gob was a self-educated turkey. He studied statistics for fun. His quarters were a bit tight at the farm, but the upside was that for 140 days straight he had been fed a nice grain meal. And as he got older, the meals got bigger. Based on this history and his statistical models, Gob had very high expectations of receiving another nice meal tomorrow from Farmer Joe, who obviously had nothing but his best interests at heart. Unfortunately, Gob would not be fed tomorrow; something different would occur. The next day was when all of Farmer Joe’s turkeys would be harvested in order to ship them to the grocery stores on time for Thanksgiving.
We can learn a lot from the past, but perhaps not as much as we think. Gob would attest to that, if he was still around. And mathematical models and statistical forecasts can lead us to be more certain of the future than we should be.
Forecasting is based on history. Many supply chain and financial forecasting models are based on bell curves, with the tails of the curve representing highly unlikely events that can be “safely” ignored, at least for forecasting purposes. But a little examination of history will show that rude surprises, like the global financial meltdown we recently experienced, can intrude. All forecasts are thrown out the window when such catastrophic events occur. That is why companies should apply supply chain risk management to product forecasts and not just to factories, suppliers, and IT systems.
Companies often develop contingency plans for facilities. Take a warehouse as an example. While you would be hard pressed to forecast exactly what event would shut down a warehouse (tornado, flood, gas leak, etc.), companies with good supply chain risk management practices have a contingency plan in place just in case. If this warehouse goes down, for whatever reason, how will we evacuate the workers? Where will we shift the work? How will we determine how much inventory to save from the warehouse? How will we shift that inventory to a different warehouse? And who has responsibility for each of the many tasks associated with the contingency plan?
Contingency plans surrounding facilities are very common. Less common are contingency plans surrounding inaccurate forecasts. Companies also need proactive plans in this area. At what point will a production line be closed down? When sales for a product family are down 6 percent? 8 percent? 10 percent?
At what point will the whole factory be shut? When sales are down 15 percent? 20 percent? If the factory is shut down, how much of the residual work gets shifted to other factories in the network? Which factories get which product lines? And, once again, who exactly is responsible for doing what?
It is not always that forecasts will be significantly too low. A key competitor, for example, could do something stupid that could benefit your company. Thus, you should have contingency plans in place for demand spikes as well. Unfortunately, downside shocks are much more common, and tend to be larger in scope, than very large upside events. Also, if there is a large amount of unexpected demand, a company can always raise prices to balance demand with supply.
The key point is that by having proactive plans companies can react more quickly to these kinds forecast error events. Having the plan in place prevents companies from dithering and can save (or make) them a whole lot of money. This kind of planning can separate turkeys slotted for slaughter from eagles prepared to soar.