Improving the Accuracy of Judgmental Adjustments to Statistical Baseline Forecasts
Published 1 February 2008
FORECASTING SUCCESS STORY: Improving the Accuracy of Judgmental Adjustments to Statistical Baseline Forecasts
EPSRC grant to analyse forecasting at McBride, Heinz, Wilkinson, and Cow & Gate
Contrary to common belief, not all expert adjustments routinely increase forecasting accuracy, as research analysing company data over a time period of three years has shown. Small adjustments can impair accuracy and up to a third of adjustments are made in the wrong direction! The findings have let to the development of a prototype software add-on that is designed to effectively incorporate expert judgment into a statistical forecast, the most common approach used in S&OP for supply chain forecasting. Research has shown that the support provided by conventional ‘notes’ systems (occasionally available in commercial systems) is limited, and that has considerable potential for improvements exists. The consequences of a successful implementation should be major improvements in accuracy.
For details of these, please contact Prof. Robert Fildes.
