Uber Freight: Assessment and Determination of Optimal Design Features for a Drop Trailer Service Offering and Network

Abstract

The trucking industry hauled 72.5% of all freight transported in the U.S. and served an essential function in transporting cargo nationwide from one place to another. Recently, the industry has suffered from significant disruption, such as a shortage of drivers and carriers. One of the leading causes of these issues is the lengthy detention time in most operational activities, such as loading and unloading time at the warehouse. Previous research recognized that the drop trailer offering serves as an effective solution for reducing the waiting time at warehouses and improving the on-time delivery rate. When it comes to our sponsoring company - Uber Freight, this type of service is still nascent, with several strategic questions unanswered. Specifically, two of the most crucial key research questions are 1) where it should expand its drop trailer service 2) what load requirement and network characteristics are best serviced with a drop trailer. Our capstone project first deployed the K-Means clustering method to address these questions to uncover the underlying pattern and key network characteristics of states that have successfully implemented the drop trailer service. The result showed that Illinois, Indiana, and Florida possess the highest feature similarity with those states and hence, are recommended for Uber Freight to introduce drop trailer service. Our project deployed a CART decision tree to decompose the critical features from our cluster results that provide a structured recommendation for drop trailer implementation to answer the second question above. The analysis indicated four features necessary for a Drop offering to be favourable compared to live loading dry-van offering. These four features lay out two sets of market conditions with their strategic consideration for Uber Freight to implement drop trailer in the future.