The rapid growth of urbanization intensifies the need for efficient passenger transportation, making public bus transport efficiency analysis crucial for smart city development. This article applies a Simar-Wilson Data Envelopment Analysis with bias correction to 37 Brazilian municipalities (2022–2024), followed by Beta and Tobit regressions to identify technical efficiency determinants in a unified framework. No municipality achieved perfect technical efficiency, with substantial performance gaps: CRS efficiency averaged 0.597, VRS efficiency 0.824, and Scale Efficiency 0.712. The 22.7% VRS-CRS difference indicates scale constraints represent more critical limitations than technical inefficiencies. Counterintuitively, municipalities implementing conventional “best practices” (user satisfaction surveys, origin-destination studies, bus corridors) exhibit lower efficiency, suggesting these practices might reflect reactive responses to existing problems rather than proactive drivers. São Paulo and Rio de Janeiro represent notable exceptions successfully combining best practices with high performance. The findings emphasize the importance of scale-appropriate policy adoption, where medium-sized municipalities should not implement metropolitan-scale solutions unsuited to their contexts. The main contribution demonstrating that practice effectiveness depends on implementation context, scale appropriateness, and organizational capacity rather than universal adoption of best practices.