Background:Retail client has over 90,000 SKUs in over 2000 stores. Client wants to build an automated pricing tool which would set optimal price points for each of the SKUs in order to maximize revenue, units sold & margins. Current pricing decisions were all intuitive & gut feel based. Through this project the client wants to leverage advanced statistical methods in pricing decisions.

Approach:Price Optimization Framework was designed, which involved the following major steps – a) Historical data analysis, b) Statistical model building, c) Price recommendation testing on a market, d) Feedback, e) Model fine tuning, f) Country wide roll out. Multiple price models were built at category, sub category & SKU levels. Regression based sophisticated statistical models were built, which accounted for price elasticity, cross-price elasticity, cannibalization, attachments, threshold price points, price gaps etc.

Impact: Client used the model on a particular sub category & saw significant lift in sales, units sold & margins.