Predictors of response, such as the ER assay and clonogenic assay, require a comprehensive clinical trial strategy. While their potential is exciting, their premature or inappropriate usage will drive up health care costs without the requisite benefit being accumplished. These assays must be demonstrated to be effective in properly designed, executed, and analyzed prospective clinical trials before they are recommended to the general population. Unfortunately, both assays have been viewed as being successful on the basis of retrospective studies. A positive retrospective study should only be viewed as a justification for the cost and effort of a large-scale well-designed prospective study. A cost-benefit analysis must be made to elucidate which subsets of patients have their therapy actually changed as a result of the assay and whether this subset actually has a better end-result which justifies the expense to the entire population. Such an analysis may also highlight subsets who do not need the assay, since it would not change their therapy. In recent years, cancer treatment has been plagued by premature claims of success for new therapies based on premature analysis of preliminary data. We should not make the same mistake with predictors of response.