Hierarchical forecasting: issues and use guidelines

In order to provide the appropriate demand forecast information given various managerial levels and functional disciplines within organizations, reliance on family‐based forecasting is increasing. The family‐based approach, sometimes referred to as hierarchical forecasting (HF), is based on a strategy of aggregating items into families. HF systems are capable of providing forecasts for items and their respective families. The objectives of HF systems, include improved forecast performance and a reduction in the overall forecasting burden. To date, several studies have offered practical guidelines for the structural design of HF systems. The primary purpose of this paper is to summarize these guidelines. First, an explanation of the HF process is provided. In this explanation, important system parameters and strategic choices, which allow for the custom configuration of HF systems are identified. Second, the relevant family‐based forecast research is reviewed. The important issues addressed and the conclusions presented in this research are identified. Third, practical guidelines regarding the use of a HF approach that have been reported in the research literature are clearly delineated. With much still unknown regarding the performance impact of various system parameter and strategic process choices, the paper concludes with suggestions for future research.