USING MARKET-LEVEL DATA TO UNDERSTAND PROMOTION EFFECTS IN A NONLINEAR MODEL

The authors show analytically, empirically, and numerically through si mulation that the estimated effects from linearly aggregated market-le vel data differ substantially from comparable effects that are obtaine d from store-level data. The magnitude of this difference renders mark et-level data largely unsuitable for econometric modeling, unless the marketing manager compensates for the bias that results from the incom patible aggregation. The authors introduce a new approach, a relativel y simple debiasing procedure derived from simulated data, They show th at this debiasing approach results in substantially improved parameter estimates. They illustrate the value of the procedure by applying ii to scanner data for powdered detergents and comparing the debiased par ameter estimates to results obtained from store-level da la and an alt ernative aggregation method that maintains homogeneity for selected pr omotional activities.