The current study investigated the impact of an 8-module internet-based cognitive behaviour therapy for perfectionism (ICBT-P) across a variety of perfectionism subscales. Undergraduate students who identified as having a problem with perfectionism were randomized to receive the intervention (n = 41), and were free to choose the number of treatment modules they completed over a 4-week period, while the control group (N = 48) received access to treatment 8 weeks post-randomisation. Secondary measures included depression, anxiety, stress, body image and self-compassion. Assessments occurred at baseline, 2-, 4- and 8-week time points. A mean of 3.12 (SD = 2.67) modules were completed; 7 participants (17%) completed none and 6 (15%) completed all. Linear mixed modelling (with baseline observation included as a covariate) showed significant Bonferroni-adjusted post-hoc between-group differences for 5 of the 6 perfectionism measures, favouring the intervention group; the most robust between group effect sizes were for the Concern over Mistakes (d = -0.82), High Standards (d = -0.69), and Perfectionistic Standards (d = -0.47) subscales. There were no between-group differences for our secondary measures. ICBT-P was found to be an effective intervention for reducing different components of perfectionism compared to a control group. The relatively low use of modules may have contributed to a lack of effect on secondary measures. Australian New Zealand Clinical Trials Registry (ANZCTR) Trial Number: ACTRN12620000562976