Description: We examine some unexpected epistemological conflicts that arise at the interfaces between ecological science, the ecosystem services framework, policy and industry. We use an example from our own research to motivate and illustrate our main arguments, while also reviewing standard approaches to ecological science using the ecosystem services framework. While we agree that the ecosystem services framework has benefits in its industrial applications because it may force economic decision makers to consider a broader range of costs and benefits than they would do otherwise, we find that many alignments of ecology with the ecosystem services framework are asking questions that are irrelevant to real-world applications, and generating data that does not serve real-world applications. We attempt to clarify why these problems arise and how to avoid them. We urge fellow ecologists to reflect on the kind of research that can lead to both scientific advances and applied relevance to society. In our view, traditional empirical approaches at landscape scales or with place-based emphases are necessary to provide applied knowledge for problem solving, which is needed once decision makers identify risks to ecosystem services. We conclude that the ecosystem services framework is a good policy tool when applied to decision-making contexts, but not a good theory either of social valuation or ecological interactions, and should not be treated as one.