Behavioral Drivers of Routing Decisions: Evidence from Restaurant Table Assignment

We first theoretically identify the factors that may impact individuals’ routing decisions before empirically examining a large operational dataset in a casual restaurant setting. Analytical models have identified various routing algorithms for service operations management. Although each model may offer advantages over others, they all make a key assumption ‐ decision makers will actually follow the algorithms, if implemented. However, in many settings routing is not done by a computer that is programmed, but instead by a human. People make routing decisions at their own discretion which may hurt or help system performance. We analyze granular transaction data to examine how hosts revise a given routing rule when seating customers. Thereafter, we empirically analyze the effect of the dispersion of table assignments on restaurant performance, and estimate the counterfactual sales impact of adopting an alternative routing priority. Our setting instructs its hosts to follow a round‐robin rule to assign tasks because it ensures fairness and smooths work flow. We find that hosts assign more incoming parties than the round‐robin rule suggests to those waiters who have low contemporaneous workload or high speed skills. The prioritization of high speed skill waiters increases with higher levels of demand. In addition, we show an inverted‐U‐shaped relationship between the inequality of table assignments (measured in terms of the Gini Coefficient of the numbers of tables assigned to each waiter during the same hour) and total sales. Our results suggest that properly adjusting the round‐robin rule is productive; however, too much deviation lowers performance. Our paper empirically highlights the value of routing decisions and front‐line personnel, such as the hosts in our context.