Key factors for instant logistics platforms to attract and retain couriers: An analysis based on online comments

While providing “last-mile” urban delivery services, instant logistics couriers (ILCs) have garnered considerable attention. This study aims to develop an influencing factors model of ILCs' willingness to join and retain on instant logistics platforms (ILPs), which can help ILP enterprises improve couriers' work motivation and reduce their turnover rate. Combining machine learning and regression analysis, this paper analyzes the online comments posted by couriers in order to provide enterprises and governments with new strategies for managing gig workers. Based on the sentiment analysis of couriers' comments, positive and negative Latent Dirichlet Allocation (LDA) topic models are developed to identify the factors that influence ILCs' job satisfaction. Then the significance of each factor on couriers' positive/negative sentiment is analyzed by a binary logistic regression model. Three positive and four negative factors are identified, with expected income having the most significant positive effect, followed by self-actualization and convenience function. In addition, unfriendly interaction with collaborators has the greatest negative impact on couriers' work motivation. An inadequate assessment mechanism also discourages couriers' participation. Similarly, entry process training and software quality are also negative factors affecting their participation behavior.