OBJECTIVES: Individualized use of available treatments – i.e., matching patients to treatments based on biomarkers or other characteristics -- holds great promise for improving health outcomes and cost-effectiveness.  However, significant resources are required to identify opportunities for such improvements and to change practice when appropriate.  When prioritizing these investments, it is important to ask “what is the maximum benefit that could be achieved via individualization?”  We developed methods to estimate this maximum possible benefit from the results of standard (non-individualized) randomized trials.   METHODS:  Previous methodological work has allowed researchers to estimate the efficiency frontier for individualized treatment assignment – i.e. the best population health outcomes that could be achieved with treatment assignments based on available biomarkers.  We derived estimators for an upper bound on this efficiency frontier. This upper bound describes the maximum benefit that could be achieved with treatment assignments based on an optimal yet unobserved biomarker.  Methods are illustrated with simulated and real clinical trial data examples. RESULTS: Upper bounds were found to depend on the distribution of outcomes in each treatment arm and on their associations with available baseline biomarkers.  In general, the stronger the association between outcomes and baseline biomarkers, the tighter the upper bound on the value of individualized care.  Examples are given in which tightened bounds are over 40% lower than bounds that do not incorporate baseline characteristics.  Examples are also given in which upper bounds suggest substantial room for improvement and, separately, in which available biomarkers nearly achieve the upper bound, indicating limited opportunity for further improvement. CONCLUSIONS: The results of existing randomized clinical trials can be used to estimate the maximum possible benefit achievable by individualizing treatment assignment.  Such findings can help benchmark the performance of available biomarkers and prioritize investment in new biomarker discovery.