Response related outcomes, when appropriately defined, play a critical role in assessing treatment effects in clinical trials. Rather than relying solely on conventional binary response indicators, a rigorous evaluation should leverage the full spectrum of response based data to more completely characterize treatment benefit and its temporal profile. We propose a comprehensive framework for response assessment that integrates multiple complementary estimands: the overall response rate or response at time of interest; the probability of being in response (PBIR) to characterize the temporal profile of response; the restricted mean duration of response (RMDOR) for unconditional analysis and inference of response durability; and the restricted duration of response (RDOR) to estimate the joint response–progression distribution and address dependent censoring. A central emphasis of this framework is the use of unconditional duration of response estimands to ensure valid statistical inference and clinically interpretable results. These methods are illustrated through simulations and case studies in oncology and rare diseases. We also discuss their potential application within the FDA’s Project FrontRunner program.