An extended Huff-model for robustly benchmarking and predicting retail network performance

AbstractThis study proposes a modi ed Hu model that takes directly into accountspatial competition between stores of the same brand, brand attraction basedon actual brand performance and spatially variable substitution. The modeluses only publicly available or easily acquirable data as input, whereas modeloutput is extensively validated on various levels. These levels include com-parison of modeled and real market shares on block, store and brand levelfor the Belgian food market. Results show that multi-objective optimizationof model parameters yields comparable results on block level to other modelsin the literature but improved results on store and brand levels, thereby en-suring model robustness. This robustness also enables the application of themodel for various business purposes as store location determination, leaetdistribution optimization, store and store concept benchmarking, withoutloss of spatial generality.Keywords:Hu model, retail management, spatial competition, multi-objectiveoptimization, store benchmarking, turnover prediction1. IntroductionTo monitor operational performance, retailers rely more and more onobjective store benchmarks. Benchmarks are objective in a way that theyquantify internal and external inuences on store performance (store size,brand, competition, geodemographic characteristics of consumers, etc.) toobtain a measure indicating the performance of the management. The more ne-grained such store benchmark is, based on for instance loyalty card in-formation, the more targeted improvement actions can be de ned. A storebenchmark on a ne-grained block level is therefore more valuable than abenchmark on an aggregate store level for de ning and monitoring the im-