Abstract: Techniques of dynamic pricing for physical stores are described. A server handling purchase transactions of multiple physical stores can aggregate sales data on goods and services and determine a pattern. The server can determine that sales of an item of goods or services at a particular physical store deviate from the pattern at a particular time. Upon determining that sales price of the item significantly correlates to the deviation, the server can determine a suggested price for correcting the deviation. As an experiment, the server can conduct an experiment of selling the item at the suggested price at the particular time for a given time period. The server can present results of the experiment to an operator of the physical store.