Price Simulation in Retail

Challenge

Price and marketing decisions today are based on the diverse information about customer and competitor behavior, particularly provided by online sales. Building the analytical capacity for data-driven, dynamic pricing is the central challenge faced by our client. Specifically, our client faced the problem of determining a pricing strategy that would not continuously lead to price reductions despite increased market transparency. Due to the high number of products and diverse data sources, this can only be managed through the application of statistical and machine learning models.

Approach

For simulating the impact of various pricing strategies, product, inventory, promotions, and sales data, along with external influencing factors, were consolidated by our data engineering team. To determine strategically relevant outcomes, such as price elasticity or competitor reactions to price changes, dedicated models were developed and combined in a simulation process. The models, simulation process, and dashboard were then integrated into the client's cloud infrastructure, where the models are automatically trained and made available for daily price simulations. As part of the project, we developed a comprehensive frontend that includes all models and allows users easy handling.

Result

With the price simulation tool, our client can now simulate and optimize the revenue and profit effects of different pricing strategies for individual products and selected sub-assortments. Different forecasting horizons allow for both short-term and medium-term analyses.