As an innovative, app-based, and demand-responsive mode of public transport, customized bus (CB) services have been increasingly promoted in cities around the world, but little is known about how it performs and what factors affect its performance. This study utilizes a practical subscription dataset for more than two years to assess the spatiotemporal patterns of CB and key factors driving the demand for CB over time. The results suggest: (1) CB services still exhibit regular (peak/off-peak) and recurrent travel patterns over time; (2) CB services are not only popularly used for commuting but for transport station (e.g., airport/railway station) transfer; (3) passenger loyalty, trip fare, and trip distance are the top three factors affecting the demand for CB; (4) the nonlinearities indicate that trip distance can be set at 20 km for maximizing the demand. To inform practice, long-distance commuting and airport transfer services from suburbs should be the two targeted niche markets for CB. Retaining existing customers rather than gaining new customers may be a more efficient strategy to increase the profitability of CB service. Locating stops near the neighborhoods with high-activity centres help increase service demand.