Dynamic pricing models in marketplace environments - the case of ride hailing business

Abstract

Classical economics theory proposes that with perfect information prices are efficient and will converge at the equilibrium of supply and demand curves. This is largely true for free markets. However, it is common for external agents e.g. governments via their central banks to propose and implement policies that exert an external force to influence prices. An example being changes to interest rates to control inflation, therefore control prices. In on-demand and online marketplaces, prices would normally be influenced by forces of demand and supply. But in most cases the owners of the marketplaces (equivalent to governments in the previous case) would want to maximize revenues by taking advantage of market inefficiencies in real-time or near real-time. This has contributed to the rise of dynamic pricing agents that act on market information to make adjustments to prices. Most dynamic pricing models’ objective to maximize revenues have a short time horizon. This study seeks to expand that time horizon by incorporating customer retention on the platforms/marketplaces to also maximize future profits; by using machine learning models to predict price sensitivity of demand and supply agents based on past behaviour.