The term structure of optimal operations

Inventory and capacity planning models generally take the time of sale as something that is exogenously given. For example, the story associated with the well-known newsvendor model is one of stocking for an upcoming selling season that will happen x units of time from now, where x is exogenous. In this paper, we re-visit the capacity planning decision by assuming that demand follows a stochastic process and study what happens when both the time of sale and capacity are decisions. When the selling price is fixed, our baseline case, we find that the optimal time to sell is either now or never. In contrast, when the selling price is stochastic, the optimal time to serve demand is somewhere between now and never. Thus, we link timing preference to two primary sources: uncertainty in demand and uncertainty in the selling price. Our results are useful whenever firms have considerable control over timing, such as in events when firms launch new products or in instances when there is no apparent selling season.