•Impact factors are still highly influential because a majority of authors benefit.•Changes in performance management must occur at the institutional level.•Evaluation has to take fundamental uncertainty in research into account.•Focal randomization mitigates biased selection of articles.•Scholarly diversity instead of one-dimensional rankings is supported. Publications in top journals today have a powerful influence on academic careers although there is much criticism of using journal rankings to evaluate individual articles. We ask why this practice of performance evaluation is still so influential. We suggest this is the case because a majority of authors benefit from the present system due to the extreme skewness of citation distributions. “Performance paradox” effects aggravate the problem. Three extant suggestions for reforming performance management are critically discussed. We advance a new proposal based on the insight that fundamental uncertainty is symptomatic for scholarly work. It suggests focal randomization using a rationally founded and well-orchestrated procedure.