Answers to Your Forecasting Questions

Article excerpt

Q. I am a Demand Planner, and our Demand Planning Group is about six months old and we are still working out some of the bugs in our systems and processes. We are calculating a MAPE value for each of our SKU, but we are having a difficulty getting a representative MAPE value forparts that have amonthly actual demand of zero. Since the formula requires the actual value to be in the denominator, we are unable to get a value because of the division by zero. By default, we have been using a value of 0.1, but that also has its limitations for giving accurate error values. Do you know of a better way to calculate these types of issues, or do we need to go to a different calculation to determine our monthly forecast error?

Rich Mayer

Paccar Parts

A. I understand your predicament. Under the circumstances, the best you can do is to compute MAPE of that SKU over a three or six month period instead of one month. Looks like that SKU is either not important or have an intermittent demand. Using 0.1 as a denominator won't help as it will distort its MAPE.

Q. I felt compelled to respond to your answer in the Fall issue of Journal of Business Forecasting on the effect of supply issues on forecast accuracy. People in sales and marketing are key components in creating an accurate collaborative demand forecast. If you start including factors in forecast accuracy that are outside of their realm of control, they will start to place less focus on forecast accuracy. This can have a negative effect on their future effort in both resolving forecast issues and in creating future forecasts. The psychological factors present in collaborative forecasting are often ignored in traditional operations management, but they are very important.

Furthermore, you pointed out that lack of supply is an issue that needs to be addressed by the organization and while I agree with that point, but it is outside the realm of forecasting. There are many other measures that a company can and should use to capture supply issues. It is definitely reasonable to expect the individual measuring the forecast accuracy to bring up specific issues with the people in the supply chain that are involved, but I would bet ten times out of ten that the sales representative that missed out on the sales would beat them to the punch.

Josh Kanownik

The Little Tikes Company

A. You brought up a very interesting point. I am one hundred percent with you that Sales/ Marketing people are the key people when it comes to preparing demand forecasts. In demand forecasting, supply people have no role to play, but demand forecasts can highlight the issue if there is any with the supply. If the supply cannot meet the demand, it will tell you that. I don't think I mentioned anything other that. Here we are talking about demand forecasts or unconstrained demand. But forecasts we use for production and financial planning can be affected by the supply.

Let me give you one example. I was sitting in one of the consensus meeting where I presented statistical forecasts. …