ALTERNATIVE METRICS FOR FORECASTING PERFORMANCE By Michael Gilliland Forecasts used in the real world are an amalgam of statistics, internal politics, and executive wishes … what is seen in the forecasting literature as essential may or may not be essential — more data, more sophisticated models, more sophisticated software and more management participation … we often waste resources by allowing forecasting participation of those whose political agendas are not aligned with unbiased forecasts. A primary purpose of business forecasting is to generate forecasts as accurate and unbiased as we can reasonably expect them to be, and do so as efficiently as possible. While we don’t have total control over the accuracy achieved, we can control the process used and the resources we invest. This article provides a method for establishing forecast accuracy expectations, and suggests a new metric for evaluating the effectiveness of each step and each participant in a business forecasting process. The approach used in measuring the “Forecast Value Added” (FVA) will help to streamline your process (making it more efficient) and improve your results (making forecasts more accurate). ...

From Issue: Winter 2003
(Winter 2003-2004)

image
Alternative Metrics for Forecasting Performance