Data-Driven Pricing for a New Product

Pricing a New Product with Data Before a product launch (or even after a launch), firms often have little demand information and often do not know critical information such as the market size, the willingness-to-pay distribution, or the adoption speed. The lack of information makes pricing a new product challenging, and insufficient data makes demand forecasting very difficult. This is particularly costly for new products because the current price not only affects the current revenue, but also the number of adopters who can influence future demand. In this paper, we consider a setting in which a firm can learn by observing early sales data at (different) prices over time. We propose a simple and computationally tractable pricing policy that guides price changes after introducing the product. Using mathematical proofs and computational study, we show that our method substantially outperforms existing methods even with a very few price changes during a selling season.