Research evaluation is praised as the symbol of modern quality management. We claim firstly, performance evaluations in research have higher costs than normally assumed, because the evaluated persons and institutions systematically change their behavior and develop counter strategies. Moreover, intrinsic work motivation is crowded out and undesired lock-in effects take place. Secondly, the benefits of performance evaluations are questionable. Evaluations provide too little information relevant for decision-making. In addition, they lose importance due to new forms of scientific cooperation on the internet. Thirdly, there exist superior alternatives. They consist in careful selection and supportive process coaching – and then leave individuals and research institutions to direct themselves.