ABSTRACT. Although ecosystem services are increasingly recognized as benefits people obtain from nature, we still have a poor understanding of how they actually enhance multidimensional human well-being, and how well-being is affected by ecosystem change. We develop a concept of "ecosystem service elasticity" (ES elasticity) that describes the sensitivity of human well-being to changes in ecosystems. ES Elasticity is a result of complex social and ecological dynamics and is context dependent, individually variable, and likely to demonstrate nonlinear dynamics such as thresholds and hysteresis. We present a conceptual framework that unpacks the chain of causality from ecosystem stocks through flows, goods, value, and shares to contribute to the well-being of different people. This framework builds on previous conceptualizations, but places multidimensional well-being of different people as the final element. This ultimately disaggregated approach emphasizes how different people access benefits and how benefits match their needs or aspirations.Applying this framework to case studies of individual coastal ecosystem services in East Africa illustrates a wide range of social and ecological factors that can affect ES elasticity. For example, food web and habitat dynamics affect the sensitivity of different fisheries ecosystem services to ecological change. Meanwhile high cultural significance, or lack of alternatives enhance ES elasticity, while social mechanisms that prevent access can reduce elasticity.Mapping out how chains are interlinked illustrates how different types of value and the well-being of different people are linked to each other and to common ecological stocks. We suggest that examining chains for individual ecosystem services can suggest potential interventions aimed at poverty alleviation and sustainable ecosystems while mapping out of interlinkages between chains can help to identify possible ecosystem service trade-offs and winners and losers. We discuss conceptual and practical challenges of applying such a framework and conclude on its utility as a heuristic for structuring interdisciplinary analysis of ecosystem services and human wellbeing.