The focus of this talk is to identify the underlying factors and develop an order assignment policy that can help an on-demand meal delivery service platform to grow. By analyzing transactional data obtained from an online meal delivery platform in Hangzhou (China) over a two-month period in 2015, we find empirical evidence that an “early delivery” is positively correlated with customer retention.
We also find that the negative effect on future orders associated with ``late deliveries'' is much stronger than the positive effect associated with early deliveries. Moreover, we show empirically that a driver's individual local area knowledge and prior delivery experience can reduce late deliveries significantly. Finally, we embed our empirical results into order assignment models to show that an on-demand service platform can provide better matching by incorporating the information from both the supply side (i.e., driver's local area knowledge and delivery experience) and the demand side (i.e., asymmetric impacts of early and late deliveries on customer future orders).