服務庫存意指廠商將「服務在製品」暫存在服務供應流程中的某一項價值活動中，一旦服務需求產生時，再將其立即轉化為「最終服務」，此種觀念在交通服務業中特別重要。本文收集衛星派遣計程車車隊300 位司機的時間與空間繞行資料，採用資料探勘及集群分析進行服務庫存實證研究，分析司機駕駛行為、服務庫存與營運績效間關係。研究發現四大類型服務庫存模式，其中，善於使用空中排班點的「彈性」型司機，其空間空車率與時間空車率均最低。本文說明了台灣大車隊如何藉由引導未載客司機前往不同「空中排班點」待命，將閒置的服務產能暫存特定區域之中；待一旦需求瞬間大量出現，便可快速派遣暫存的空車前往服務，滿足即時需求。文末提出理論與實務意涵。

The term “service inventory” implies that a firm keeps a “service in process” within a service-based supply chain as a form of temporary stock. The “temporary service stock” belongs to a certain value activity within the service supply chain; once the service is demanded by a customer, the service provider will deploy temporary service as “the final service”. Due to the lack of empirical studies regarding this concept, we conducted research by way of a quantitative approach, using the data mining and cluster analysis methodologies to identify and analyze the relationship between service inventory and operational performance. By using time-space air-queuing data from a group of 300 taxi drivers provided from the Taiwan Taxi Fleet group’s satellite system, we identified four categories of taxi drivers’ service inventory patterns: “Customized”, “High-Speed”, “Flexibility” and “Routine”, respectively. Furthermore, we investigated the relationships between these four patterns and their performance, with the end result showing that most drivers of the fleet qualify in the Flexibility category, efficiently overcoming issues in service perishability and unstockability by utilizing the “on-the-air queuing system” for the taxi service. This study concludes with theoretical and practical implications for academia and practitioner reference.