Policy optimization, a variant of The Input-Environment-Output's (I-E-O's) environmental optimization, has found traction on the edges of and outside of the discipline of higher education. It has spawned an entire industry of vendors (e.g., EAB, Starfish, and the myriad vendors specific to functional areas) selling student success solutions, and it generates limitless revisions of state performance-based funding standards (e.g., Florida) to optimize public expenditures. Perhaps most insidiously, this common sense logic insists that all campus actors properly desire to optimize. Optimization is only possible when know-how meets a campus culture of data. One National Institute for Learning Outcomes Assessment report names this combination of knowledge and desire an "improvement agenda." Paradoxically, policies that aim to optimize student success are the very engine that creates the problem that feeds the need for campuses to desire data in order to optimize success. In other words, the very policies that aim to solve the problem of student success create the problem of student success, as insofar as these policies define student success, they also by definition define student failure. As these policies create the boundaries of student success that also create the boundaries of student failure, they place these failures outside the reach of current optimization efforts, and the reason for policy interventions to continue. To think and act outside of optimization, trust and actions should be placed in values systems that make optimization the problem.