Customer Herding in Services with Waiting Costs

Abstract

We explore customer choice behavior when they face a choice between service providers of unknown service value. Customers arrive according to a Poisson process to the market. Service times are exponentially distributed with the same rate at each service provider. Both the service providers are capacitated (finite buffers), and the customers incur waiting costs when waiting for service. Customers observe decisions made by other customers in the form of queue lengths before making their choice. If more customers choose the same service, longer queues will form at the service. Therefore, the arriving customers infer service value is better at the service with the longer queue and would prefer to join the longer queue. But customers also have to suffer additional waiting costs in the longer queue before consuming the service. Furthermore, they might not even get the service at longer queue sometimes because the waiting space is full. We characterize the equilibrium queue-joining behavior of arriving customers under the presence of such positive and negative externalities.

In addition to the above-mentioned externalities, customer choice is influenced by heterogeneity in the customer information. When endowed with noisy private signals, customers might ignore their own private information and join the longer queue. We find that even when the waiting costs are positive, it might be rational for customers to join the longer queue when the total arrivals to the market is not very high compared to total service rate. Therefore, market shares of high quality firms improve under positive externalities when the arrival rates are low or when the service rates are fast. Finally, we show that the most informed customers in a market do not herd, even if their signal were quite poor, as long as they had better signals than other customers in the market. Thus, we demonstrate that heterogeneity in customer information is a bigger driver of herding behavior than queue lengths per se. Agents with only marginally better and imperfect information can trigger other agents in the market to herd.

Keywords: Customer Herding Behavior, Waiting Costs, Service Capacity, Uncertain Service Value

JEL Classification: D62, D82, D83