Investigating the cross-sales effect of product associations

Abstract

In this research, we use the framework of association rule discovery to find 1,350 interesting product associations between two sku’s. Using multivariate time series techniques, we successively simulate a price promotion in both products and measure the impact on the sales of the associated product. This approach allows us to model both the short run and long run cross-sales effect. For both complement and substitute relationships, we investigate the moderating effect of several covariates on the size of the cross-sales effect.