Achieving perfect power proportionality in current mobile devices is not enough to prevent users from running out of battery. Given a limited power budget, we need to control active power usage, and there needs to be a prioritization of activities. In the late 1990s, Flinn and Satyanarayanan showed signicant energy savings using a concept of data fidelity to drive mobile application adaptation, informed by the battery lifetime desired by the user and the OS’s evaluation of energy supply and demand. In this paper we revisit and expand this approach, recognizing that with current hardware there are even higher potential savings, and that increased diversity in applications, devices, and user preferences requires a new way to involve the user to maximize their utility. We propose Application Modes, a new abstraction and a narrow interface between applications and the OS that allows for a separation of concerns between the application, the OS, and the user. Application Modes are well suited to eliciting user preferences when these depend on multiple dimensions, and can vary between users, time, and context. Applications declare modes – bundles of functionality for graceful degradation when resourcelimited. e OS uses these modes as the granularity at which to profile and predict energy usage, without having to understand their semantics. It can combine these predictions with application-provided descriptions, exposing to the user only the high-level trade-os that they need to know about, between battery lifetime and functionality.