Different forms of psychotherapy are effective for cluster C personality disorders, but we know less about what in-session processes promote change. Contrasting successful and unsuccessful cases may elucidate processes that facilitate or impede outcome and offer suggestions for clinical practice and future research. In this exploratory outcome-process study, 10 successful and 10 unsuccessful cases were selected from a randomized trial comparing cognitive therapy and short-term psychodynamic psychotherapy for cluster C personality disorders. Videotaped sessions were rated with the Psychotherapy Process Q-Set (PQS). The treatments were compared in terms of which PQS items differentiated successful and unsuccessful cases, as well as their resemblance with PQS prototypes of "ideal treatments." Therapists behavior in early sessions was also explored. Results indicate that successful cases in our sample were characterized by a more active and engaged patient. In contrast, unsuccessful cases were characterized by a more directive or "controlling" therapist stance. Correlations with PQS prototypes were moderate to strong in both successful and unsuccessful cases, suggesting that optimal and suboptimal interpersonal processes may be independent of adherence to particular treatments. Exploration of therapist behaviors in early sessions indicated that therapists were more likely to adjust their way of working in the successful cases. Our result suggests that patient engagement and therapists early efforts to improve the therapy relationship may be pivotal for successful outcome, whereas therapist controlling behavior may obstruct the treatment process, regardless of therapy model used. The impact of these in-session processes should be examined more closely in larger samples in future studies.