Objective The present study was a replication and extension of prior work (Stulz, Lutz, Leach, Lucock, & Barkham, ) that identified multiple groups of clients in treatment with high‐symptom severity and markedly different recovery trajectories (rapid/early response vs. little or no response). Method Using data collected through repeated administrations of the Depression subscale of the Treatment Outcome Package (n = 147), growth mixture modeling was employed to determine whether clients fell into discrete groups of response trajectories during 15 sessions of psychotherapy. Additionally, logistic regressions were conducted to assess possible predictors of group membership. Results Three separate groups of treatment responders were identified: 2 high‐symptom groups–rapid responders and nonresponders–and 1 low‐symptom group of nonresponders. Elevated social conflict and suicidality predicted increased likelihood of membership in the high‐symptom nonresponder group. Increased feelings of interpersonal hostility and better sexual functioning predicted increased likelihood of membership in the rapid responder group. Conclusion Replication of earlier results provides further evidence for the usefulness of modeling change during psychotherapy using multiple trajectories. Predictors of group membership indicate the influence of functional impairment on recovery, and support the importance of multidimensional measurement of client problems.