Context: Crowd logistics is a novel shipping concept where delivery operations are carried out by employing existing vehicle capacity and drivers from the crowd, relying on their planned tours, thereby offering potential for economic, social, and environmental benefits. Despite the promise of this new logistics model, little is known about its actual functioning, performance, and impact. In recent years, goods delivery systems are under increasing pressure from e-commerce, coupled with increasing consumer expectations for delivery performance (e.g. traceability, same-day delivery, home-delivery). At the same time, new opportunities are unlocked by technological innovation, including tracking technology and increased use of connected mobile technology by customers. These innovations have fostered the creation of new enterprises and business models in the logistics sector. Crowd-shipping presents a number of specific features and research challenges that call for rigorous empirical investigation. While crowd-logistics operations tend to be affordable and convenient, the performance is dependent on non-professional (occasional) couriers that do not offer traditional guarantees. The lack of locally developed networks of senders and drivers can cause deliveries to be delayed or even prohibit matching. Another unique feature that warrants further study is the interaction among senders and drivers that strongly impact the efficacy of the bidding and shipment process managed on the virtual platform. The flexibility and scalability of the system represent both a strength and a weakness of the system. In summary, crowd-shipping has the potential to have a positive impact on the shipping industry and society, but several obstacles curb the growth of the sector and the development of the crowd-shipping companies which sometimes have difficulty reaching a critical mass of users. There is a need for fundamental research to understand the functioning and performance of these emerging shipping systems. In particular, research on user behavior, network formation and overall reliability and performance of crowd-shipping will contribute to the development of the system. Objective and Organization: The goal of this dissertation is to examine the properties and performance of crowd-shipping systems by investigating crowd-shipping users’ behavior. The research benefits from the access to numerous sources of data, including a large-scale database with real crowd-shipping operations. The work is divided into four main streams, each focusing on a specific part of the crowd-shipping system. First, we focus on the behavior of senders on the platform. As the features of crowd-shipping services are different from those of the traditional shipping industry, it is important to know who the users are and which features of the system impact their decision to adopt crowd-shipping. Through a binomial logit model and a structural equation model, we define the characteristics of the crowd-shipping users which distinguishes them from the non-users by looking at their socioeconomic and demographic information, their personal motivations, as well as the built-environment. Then, we build a set of discrete choice models to highlight the factors which influence the acceptability and preferences for crowd-shipping attributes, and assess the role of context and experience effects. Second, we study the interactions between crowd-shipping senders and couriers. Prior to a delivery, senders and drivers communicate on the online platform in order to reach an agreement for the delivery. Content and sentiment analysis techniques are used to study the messages exchanged between senders and drivers with the goal of predicting the outcome of these conversations: can the message exchanges reveal the path to finding an agreement leading to a successful delivery? In addition, we use descriptive network analysis techniques to examine the various interactions between the crowd-shipping users in a multilayer network. The network perspective provides an opportunity to visualize the functioning of the crowd-shipping platform and to study the behavior of its agents in the network via the formation of ties. Third, we pursue the application of network analysis tools to focus, in turn, on driver’s behavior. Specifically, we develop an exponential random graph model applied to the network of drivers to highlight homophily and spatial proximity in driver’s bidding behavior. Fourth and finally, we investigate the overall performance of the system. We develop a series of random forest models to define indexes which evaluate the outcome of a specific package delivery request as a function of package and broader built environment features. Contribution: This dissertation research contributes to the state-of-the-art by providing unique insights on crowd-shipping users’ behavior and the performance of the system, using various innovative modeling techniques. It establishes a foundational portrait of crowd-shipping users studying their behavior, attitudes, motivations, characteristics, and interactions with other users of the system. The analysis relies on two unique sets of data: an online survey which studies the general public’s crowd-shipping use and acceptance, and a large-scale crowd-shipping company operational database. The latter contains the list of all the users of a leading U.S. crowd-shipping platform and their information, as well as the records of each shipment request and the text communication exchanged between the users. Findings of the dissertation benefit academic research and practice in several ways. They offer insights for logistics researchers to define more realistic assumptions about the users from the crowd in their models and get more accurate results. Results would also benefit the crowd-shipping companies who would have a better understanding of the needs of their users and would know how to address them. Overall, the findings from this dissertation brings evidence of the main drivers of crowd-shipping performance, with detailed information, including elasticities for a range of factors, that can guide the fundamental understanding, as well as regulation and policy stimulation of crowd-shipping operations.

Creator

• Punel, Aymeric

DOI

• https://doi.org/10.21985/n2-cve8-tb10

Subject

• Transportation

Language

• en

Alternate Identifier

• http://dissertations.umi.com/northwestern:14592
• etdadmin_upload_660772

Keyword

• Discrete Choice Modeling
• User's Behavior
• Sharing Economy
• Crowd-shipping
• Social Network Analysis
• Urban Freight Delivery

Date created

• 2019-01-01

Resource type

• Dissertation

Rights statement

• In Copyright