This article elucidates the premises and limited meaning of medical futility in order to formulate an ethically meaningful definition of the term, that is, a medical intervention's inability to deliver the benefit for which it is designed. It uses this definition to show the two ways an intervention could become medically futile, to recommend an even more limited usage of medical futility, and to explain why an intervention need not be futile in order to be withdrawn over patient-based objections. If an intervention retains some benefit, then patients or surrogates might legitimately consider that benefit in their case and request the intervention. Physicians might still be justified in declining it on the grounds that the burdens greatly outweigh the benefits, but not on the grounds of futility. Finally, the article uses bioethics research and healthcare litigation to clarify the meaning of futility in practice and recommends alternative language when possible.