The Forecasting Newsvendor: Benefits of Improved Demand Forecasts When Acquisition Costs are Increasing

Abstract

We investigate the supply strategy of a firm that faces uncertain seasonal demand. While additional demand information becomes available in the form of signals throughout a finite planning horizon, it also becomes more expensive to acquire the product. In this setting, we explore the tradeoff between increasing acquisition cost and benefits of improved demand forecast. First, we show that the optimal replenishment policy is a state-dependent base-stock policy, where the state represents a statistic of the observed signals. Then, we characterize the base stock policy and show that the optimal base stock level is a linear function of the last updated mean demand. We also discuss the impact of fixed ordering and information acquisition costs on the firm's supply strategy. Finally, using a special case of our main model, we characterize the conditions under which a responsive, frequent ordering strategy with information updating is valuable.

Keywords: newsvendor model, inventory management, forecasting, forecast updating