As the rapid growth of urban e-commerce increases the volume of last-mile deliveries, logistics service providers have difficulty in meeting the demand of on-demand consumer requests. This increase in demand challenges traditional delivery, with some parcels becoming disproportionately costly to deliver to their destinations. To address this, we introduce a cost-based outlier parcel selection mechanism that identifies parcels with a high negative impact on the marginal delivery costs. These outlier parcels are then eliminated from their tours and outsourced to a crowdshipping market, where individuals combine the delivery task with their already planned trips. We use unique data on delivery tours of six service providers for the province of South Holland in the Netherlands. The cost-based decision rule for identifying outlier parcels results in a low proportion of outsourcing to the crowdshipping market compared to earlier literature. We identify only about 1 % of the total parcel demand as outliers across all carriers combined. Of these outlier parcels, the proportion selected for crowdshipping based on their cost efficiency ranges from 42.78 % to 3 %, depending on the scenario. While crowdshipping provides a viable solution for handling a small portion of last-mile deliveries, its environmental and economic sustainability is restricted by factors such as compensation rates and the delivery mode used. This study demonstrates that outsourcing high-cost outlier parcels to crowdshipping can be cost-efficient and reduce emissions of last-mile logistics companies; however, the proportion of these parcels is very small, limiting the overall impact on sustainability.

Thanks to growing online shopping, last mile logistics is becoming a more relevant problem for cities due to its negative impacts, such as congestion and environmental problems. In this research, one of the urban freight transport services aiming to tackle these externalities is analysed: crowdshipping. Crowdshipping is a service where the package is delivered via a traveller who is already making an unrelated trip. Trust is a key concept affecting the adoption of crowdshipping yet to be explicitly investigated . Thereby, this research aims to explore the effect of trust and examine how users' adoption can be achieved. To analyse the stated gap, a stated choice experiment is conducted to test the effect of travel time, travel cost, track and trace, insurance, damage and reputation. A Mediation Choice Model is applied to explore how the relevant attributes would impact trust towards the service adoption. Based on the findings, the direct effects of all the selected attributes were found significant, except for tracking and tracing. Regarding the indirect effects, all the main attributes were statistically significant, meaning that trust has a mediating effect on the adoption of the service. Additionally, the heterogeneity in preferences is explored through a Latent Class Choice Model resulting in a two-class latent model, crowdshipping sceptics and crowdshipping enthusiasts, the latter being more likely to be composed of younger women. Although enthusiasts consider more features while opting for crowdshipping, the delivery company's reputation and delivery cost were the most important factors in crowdshipping service adoption.