People with psychosis benefit enormously from Cognitive Behavioural Therapy for Insomnia (CBTI), although some variability exists in treatment outcomes. While recent efforts have focused on profiling sleep at treatment initiation, an alternative methodological approach involves using treatment response as a starting point to better understand what constitutes an effective treatment. This study used Grade of Membership analysis (GoM) to estimate the occurrence of unique treatment outcomes and associated patient characteristics. Outcome measures included changes in sleep (self-reported latency, efficiency, duration, quality), functional outcomes (daytime dysfunction, negative mood, need for sleep medication) and treatment goal, collected in 50 individuals with a psychotic disorder and insomnia who underwent CBT-I treatment. Three distinct profiles were identified: (1) Strong responders, who met their treatment goals and showed broad improvements in both sleep and functional domains; (2) Partial responders, who showed sleep improvements (particularly in total sleep time), without noticeable gains in function, and who were predominantly female; and (3) Non-responders who showed little treatment response and failed to meet their treatment goals. This group was also more likely to have greater psychopathology (high levels of negative mood and psychotic symptoms, and antipsychotic medication dosage). These findings suggest that (i) CBT-I can serve to improve sleep and daytime function (although sleep can improve independently from function), (ii) client-elicited treatment goals are a key predictor of CBT-I response, (iii) other important variables associated with treatment response include sleep profile, clinical acuity, and sex.