Behavioral Heterogeneity in Mobility as a Service Bundle Choice: A Latent Class Logit Model Approach in Beijing

Mobility as a Service (MaaS) is a key approach to advancing sustainable urban mobility, mainly accessible to users through bundled services via mobile platforms. However, empirical studies focusing on preference segmentation and willingness-to-pay (WTP) for bundled MaaS services remain limited, especially in rapidly urbanizing cities such as Beijing. To address this gap, this study developed a latent class logit model integrating latent psychological attitudes along with sociodemographic and travel attributes to identify latent user classes and determine key factors influencing bundle choice behavior, which subsequently provides a comprehensive perspective for understanding behavioral heterogeneity in MaaS bundle choices. Based on 485 stated preference questionnaires collected in Beijing, three distinct latent user classes were identified: potential adopters (Class 1, 29.0%); MaaS-indifferent individuals (Class 2, 58.8%); and avoiders (Class 3, 12.3%). These classes exhibit significant differences in characteristics, preferences, and WTP for the monthly bundle components. Class 1 shows a strong preference for metro ridership quotas with a WTP ¥1.01, a negative attitude and reluctant to pay ¥1.83 toward taxi mileage, and high WTP ¥84.47 for the shared function. Class 2, most representative of Beijing’s population, shows a positive preference for bus ridership with WTP ¥2.48 and is more strongly influenced by psychological attributes. Class 3 prefers “pay-as-you-go” but shows positive preferences for bus, metro and shared function in the bundle with respective WTP ¥1.90, ¥2.05, and ¥191.18. This study contributes empirical evidence of behavioral heterogeneity in MaaS adoption and offers practical implications for targeted MaaS product design and policy making.