Fildes, R A and Goodwin, P and Lawrence, M and Nikolopoulos, K (2006) Effective forecasting for supply-chain planning: an empirical evaluation and strategies for improvement. Working Paper. The Department of Management Science, Lancaster University.
Abstract
Demand forecasting is a crucial aspect of the planning process in supply-chain companies. The most common approach to forecasting demand in these companies involves the use of a simple univariate statistical method to produce a forecast and the subsequent judgmental adjustment of this by the company's demand planners to take into account market intelligence relating to any exceptional circumstances expected over the planning horizon. Based on four company case studies, which included collecting more than 12,000 forecasts and outcomes, this paper examines: i) the extent to which the judgmental adjustments led to improvements in accuracy, ii) the extent to which the adjustments were biased and inefficient, iii) the circumstances where adjustments were detrimental or beneficial, and iv) methods that could lead to greater levels of accuracy. It was found that the judgmentally adjusted forecasts were both biased and inefficient. In particular, market intelligence that was expected to have a positive impact on demand was used far less effectively than intelligence suggesting a negative impact. The paper goes on to propose a set of improvements that could be applied to the forecasting processes in the companies and to the forecasting software that is used in these processes.