Social comparisons are frequently used for self-evaluation. As a consequence, judges have to be efficient in each comparison step. Standard selection, however, is traditionally seen as an elaborate process in which judges deliberately select similar standards. We propose that often judges do not engage in such deliberate selection processes. Instead, they use routine standards--standards that have frequently been used for self-evaluation. Consistent with this assumption, Studies 1-3 demonstrate that student participants activate knowledge about their best friend--a likely routine standard--during self-evaluation. In Study 1, lexical decisions for the best friend's name are facilitated after self-evaluation. In Study 2, participants judge their best friend faster after evaluating themselves on the same dimension. In Study 3, this is even the case if the best friend is dissimilar and consequently undiagnostic. Finally, in Study 4, self-evaluations are primarily influenced by comparison information about an experimentally created routine standard.