Flexible estimation of price response functions using retail scanner data

Accommodating heterogeneity and nonlinearity in price effects for predicting brand sales and profits

“…On the other hand, researchers have proposed nonparametric regression models in order to accommodate potential nonlinearities in store sales response (e.g., Kalyanam and Shively, A C C E P T E D M A N U S C R I P T 4 1998; van Heerde et al, 2001;Steiner et al, 2007;Brezger and Steiner, 2008;Haupt et al, 2014). The empirical results of this second stream indicate that own-and cross-price effects may show complex nonlinearities which are difficult or not at all to capture by parametric models.…”

Smooth Quantile-Based Modeling of Brand Sales, Price and Promotional Effects From Retail Scanner Panels

“…for the monotonicity‐constrained version) we employ equation to obtain smooth negative first derivatives (e.g. Bollaerts et al ., ; Steiner et al ., ). We choose the knot sequences for the two B‐spline models to be equidistant, thereby reducing the problem of knot selection to specifying only the number of interior knots m .…”

“…To account for the price‐quality tier structure of the data, one could capture cross‐price effects at the tier level (e.g. Steiner et al ., ). This would yield a more parsimonious model on the one hand but, on the other hand, may affect the estimates of own‐price elasticities and/or the predictive performance of a model due to that aggregation.…”

“…We therefore explicitly utilize the panel structure of the data and propose to use a rolling window approach for prediction, which allows for the dependence between observations over time in the data. This is contrary to previous studies that have used random sampling procedures for cross‐validation (Kalyanam and Shively, ; van Heerde et al ., ; Steiner et al ., ; Brezger and Steiner, ). In particular, we assess the accuracy of our out‐of‐sample sales predictions separately with respect to both the cross‐sectional dimension (i.e.…”

“…More recently, using a penalized spline approach, Steiner et al . () and Brezger and Steiner () illustrated that nonparametric estimation of price effects can significantly improve the predictive performance of a store sales model, even when price response curves were constrained to show a monotonic shape, as suggested by economic theory (e.g. Rao, ; Hanssens et al ., ).…”