Ecosystem service research covers a challenging socio-ecological complexity and simultaneously copes with a high policy demand for decision support in sustainable resource management. This stimulates proliferation of pragmatic modeling techniques, such as the matrix model: ecosystem service supply is modeled using expert estimations per land use or land cover class. The matrix models popularity proves its main strengths (efficient, fast, accessible and adaptable), but also entails risks for scientific credibility and legitimacy of its results and ecosystem service assessments in general. Some of the main methodological critiques on the matrix model can be addressed especially by including measures of confidence, traceability, reliability, consistency and validity. This review presents recommendations and encourages these to become standard practise in future applications of the matrix model and related techniques. Additionally, we argue that an extended matrix model could provide more than only scientifically sound and politically legitimate results. It could serve as a tool to improve cooperation between natural and social sciences, experts, stakeholders and decision makers: collaborative development of the matrix model contributes to transdisciplinary ecosystem service research aimed at effective implementation and action