Applying multiple regression analysis to the forecasting of grocery store sales: an application and critical appraisal

Multiple regression analysis is applied to an analysis of the trading performance of a chain of grocery stores in the north-east of England. The performance of stores outside the conurbations is shown to be related to characteristics of the OPCS ‘urban-areas’ in which the stores are located, and a way is presented of modelling the effect of competition from within and outside these centres.

A high degree of statistical explanation is achieved, and the methodology used is defended against criticisms that can be levelled at many of other multiple regression analyses of store performance. Caution is nevertheless advised against the application of the model to forecasting the performance of new stores, although a range of applications are suggested whereby the implications of changes in the trading circumstances of existing stores might be examined.