MVVM in the Era of Modern Android: A Systematic Literature Review and Taxonomy of Architectural Trade-offs
Fabian-Enrique Suarez-Carvajal, René Lobo-Quintero, Julián Santiago Santoyo Díaz, Javier Pinzón Castellanos, Yamid Gamba-Gonzalez · CLEI electronic journal · 2026
Model-View-ViewModel (MVVM) has become a widely adopted architectural pattern in mobile application development, driven by modern programming languages and the transition toward declarative user interfaces such as Jetpack Compose. Despite its popularity, empirical evidence regarding the impact of MVVM on software quality attributes remains fragmented and often contradictory, particularly within the Android ecosystem. This paper presents a Systematic Literature Review (SLR) of 78 primary studies published between 2017 and 2025, following the guidelines of Kitchenham and Charters [1] and sourced from five major scientific databases. The review focuses on MVVM implementations in modern Android development, with contextual comparisons to iOS and cross-platform environments.The study contributes (i) a taxonomy of research intentions and adoption contexts for MVVM in mobile software development, (ii) a systematic synthesis of reported trade-offs between key quality attributes, most notably improvements in testability and maintainability versus performance and resource overhead and (iii) an analysis of architecturalchallenges associated with reactive and declarative paradigms. The results indicate a strong predominance of Android-focused studies and a technological shift toward coroutine and flow based implementations, while also revealing recurring performance and memory concerns linked to excessive data binding and observer usage. The findings suggest that, although MVVM effectively supports maintainability and testing practices, its uncritical adoption in modern mobile applications may lead to architectural degradation if performance and energy implications are not explicitly addressed.The paper concludes by outlining a research agenda focused on empirical validation in large-scale projects and the standardization of MVVM practices in declarative UI environments.