The dominance of ''ecosystem services'' as a guiding concept for environmental management - where it appears as a neutral, obvious, taken-for-granted concept - hides the fact that there are choices implicit in its framing and in its application. In other words, it is a highly political concept, and its utility depends on the arena in which it is used and what it is used for. Following a political ecology framework, and based on a literature review, bibliometric analyses, and brief examples from two tropical rainforest countries, this review investigates four moments in the construction and application of the ecosystem services idea: socio-historical (the emergence of the discourse), ontological (what knowledge does the concept allow?), scientific (difficulties in its practical application), and political (who wins, who loses?). We show how the concept is a boundary object with widespread appeal, trace the discursive and institutional context within which it gained traction, and argue that choices of scale, definition, and method in measuring ecosystem services frustrate its straightforward application. As a result, it is used in diverse ways by dif- ferent interests to justify different kinds of interventions that at times might be totally opposed. In Madagascar, the ecosystem services idea is mainly used to justify forest conservation in ways open to cri- tique for its neoliberalization of nature or disempowerment of communities. In contrast, in the Brazilian Amazon, the discourse of ecosystem services has served the agendas of traditional populations and family farm lobbies. Ecosystem services, as an idea and tool, are mobilized by diverse actors in real-life situa- tions that lead to complex, regionally particular and fundamentally political outcomes.