<p>Mobility as a Service (MaaS) aims to reduce the complexity of travel planning and execution for multimodal services. Understanding customer behaviour in real market conditions is crucial for designing and maintaining effective MaaS strategies. However, there is limited empirical data on customers actual subscription patterns, and few studies explore longitudinal behaviour in stable MaaS environments. This study narrows these knowledge gaps based on trial data collected from a MaaS trial at the University of Queensland (UQ), in Brisbane, Australia, with a focus on customers’ MaaS subscriptions during the first semester of 2023 (from 20 February to 17 June). A sequence analysis reveals that there are seven types of representative subscription patterns in customers’ longitudinal sequences. Each pattern significantly differs in terms of the temporal distribution of daily membership states, primarily around the duration of the subscription and the variation of mobility bundle choices over time. Overall, within the MaaS bundles on offer in this trial, customers are more likely to subscribe to mobility bundles only with unlimited allowances for public transport service in the long run, while only a minority of customers subscribe to multimodal mobility bundles. A multinomial logit (MNL) model based on the seven subscription patterns also confirms the important roles of customer demographics, motivations for joining the MaaS trial, built environment characteristics surrounding home locations, and trip booking characteristics, in shaping the heterogeneous subscription patterns. The findings are insightful for MaaS operators/practitioners to design bundle offerings, thus retaining customer subscriptions. The proposed research framework can also be used to construct customer market segments or to plan future bundle offerings in MaaS markets.</p>