Opaque Selling with Dynamic Pricing and Demand Learning

We study a seller managing a finite inventory of two substitutable products over a finite time horizon. To better align demand with available supply, the seller implements a novel selling mechanism: a dynamic opaque product. An opaque product is an additional option within the existing assortment, the exact identity of which is revealed to the customer immediately after purchase. The seller can dynamically adjust both the realization and the price of the opaque product based on the remaining time in the horizon and the current inventory levels. We demonstrate that the optimal opaque policy can increase revenue and reduce waste in finite-horizon inventory settings. We further observe that the opaque selling strategy can lower stockout probabilities under a wide range of parameter settings. Finally, we show that the demand distribution as a function of the opaque product’s price can be learned with an acceptable regret rate.