Price Elasticities in Retail

Challenge

Estimating product or product group-specific price elasticities is a typical use case for statistical models in the retail environment. Typically, this involves modeling daily or weekly price and sales information while controlling for other influencing factors such as promotions, seasonal factors, and competitor information. The goal of this project was to test the calculation of price elasticities at the SKU level, while controlling for various factors, particularly competitor pricing information, for our client.

Approach

After extensive preparation of the data at the receipt level, various data aggregates were created on a daily and weekly basis. These served as the basis for estimating the elasticity models. The usual approach of estimating simple linear models was modified to model a store-specific deviation from the average elasticity for each product. This ensured that the varying strength of the relationship between price and quantity sold for different stores was adequately considered in the model.

Result

As part of the elasticity calculations, valuable insights were gained into our own data assets and structures, as well as the mechanisms and relationships between our own and competitor prices. The calculated elasticities were used to test new pricing approaches and to further develop and apply the elasticity model to new markets on the client side.