The customer sets daily prices for ~1,000 customers for ~6,000 items. Due to the volume, only flat-rate margin mark-ups can be applied on the basis of customer classifications. This simplified approach ensures that some margin potential is not realized and that sales are completely lost elsewhere.

By using complex pattern recognition algorithms, we have determined precise price elasticities per customer and item. These could also be taken into account thanks to available competitor prices. In addition, external factors such as weather, vacations, public holidays, demographic information and general economic indicators play a central role.

The second solution component compares the inventory with the demand forecast. If demand is higher than inventory, a minimum price is set to avoid unnecessarily cheap sales.