Return rate prediction in e-shop saves more than 100 000 EUR per year
A large e-shop in the Czech Republic has a very large portfolio and wanted to predict which items will have a higher return rate.
We delivered a quantile regression model per each product that predicts an individual item's return rate.
The client is now able to detect problematic products and take them off the shelf early, thereby saving return costs. Cost savings of 2 500 000 CZK per year.