Motivated by recent developments aimed to address last-mile delivery challenges in rural areas of Indonesia and China, we develop a theoretical model to study how to lower participation barriers to the e-commerce market. We analyze how stage stations lower the cost of last-mile delivery by leveraging the preexisting social fabric in rural communities, including local stores, as well as social technologies such as chat services and virtual groups. Specifically, we examine two models, a decentralized model with stage stations run by independent agents, and a centralized model, similar to the one pioneered by Alibaba Taobao. We find that when the delivery cost in rural areas is high, the decentralized model can lower the participation cost and increase both platform profit and social welfare. In general, the centralized model outperforms the decentralized one. One reason is that the decentralized model suffers from a double marginalization problem. The centralized model has more flexibility, for example, in determining the pricing policies for both sellers and buyers. It could offer free service to rural customers if the participation of fresh rural customers can attract many additional sellers. However, a centralized model may not be feasible in many countries. Therefore, we explore whether the platform can implement a coordination mechanism for decentralized stage stations. We find that our proposed coordination mechanism can improve the performance of the decentralized model. Our results have important implications for how e-commerce platforms can leverage stage stations and social technologies to lower participation barriers for rural customers, thereby creating a more inclusive development model.