Previous research has shown cognitive-behavioural treatment (CBT) to be effective in reducing perfectionism. The present study investigated the efficacy of two formats of CBT for perfectionism (CBT-P), face-to-face and pure online self-help, in reducing perfectionism and associated psychological symptoms. Participants were randomly allocated to face-to-face CBT-P (n = 18), pure online self-help CBT-P (n = 16), or a waitlist control period (n = 18). There was no significant change for the waitlist group on any of the outcome measures at the end of treatment. Both the face-to-face and pure online self-help groups reported significant reductions at the end of treatment for the perfectionism variables which were maintained at the 6-month follow-up. The face-to-face group also reported significant reductions over this time in depression, anxiety, and stress, and a significant pre-post increase in self-esteem, all of which were maintained at the 6-month follow-up. In contrast, the pure online self-help group showed no significant changes on these outcomes. The face-to-face group was statistically superior to the pure online self-help group at follow-up on the perfectionism measures, concern over mistakes and personal standards. The results show promising evidence for CBT for perfectionism, especially when offered face to face, where sustained benefit across a broad range of outcomes can be expected.