Crowd-starting a Shared (Shuttle) Service with Customer Engagement

Recent developments of smart-city operations empowered by digital platforms have encouraged customer engagement to improve operational efficiency and service quality. In urban mobility, the shared shuttle service is emerging as an intermediate mode between public transport and ride-sharing services. In particular, platforms can offer new shared shuttle services based on customer suggestions in a crowd-starting manner. This type of customer engagement reveals additional information about customer needs, meanwhile poses a significant challenge in the service design for the platform—how to trade-off between service coverage that aims to share the service with more customers and service quality that aims to meet one’s personal needs. The crux of the problem is to strike a balance in this trade-off by understanding how the information from customer suggestions can be leveraged with proper optimization techniques for a better service design. To this end, we propose a service design optimization model to maximize the expected adoptions, the key objective in crowd-starting business. We first employ the preference list model to characterize how customers will respond given different service attributes and how customer suggestions are related to the responses. We then quantify the value of information from customer suggestions and investigate its relationship with the information gain. Moreover, for a data-driven implementation, we employ isotonic regression to calibrate the adoption probabilities and propose a modified approach with shrinkage for the small-data regime. Finally, we conduct a case study for the shared shuttle service and discuss practical insights, e.g., when it is beneficial to encourage customer engagement in the service design process