Background: The study examined the effect of the early working alliance on outcome in outpatient substance abuse treatment. Methods: A total of 327 clients and 33 therapists participated in the study. Data were collected in southern and western Finland in outpatient treatment units (N = 7). The dependent variables were percentage of days abstinent and client satisfaction at six-month follow-up. The independent variables were both client and therapist alliance ratings of the first and third sessions. The client's percentage of days abstinent for the month preceding treatment was used as a covariate. Intra-class correlation was used to measure between-therapist variation. Results: The main finding was that there was considerable between-therapist variation in both the frequency of clients' substance use and client satisfaction at follow-up. The client's earlier substance use frequency was a significant predictor of the substance use frequency at follow-up. Client satisfaction was significantly predicted by the client's rating of the early alliance. Conclusions: More research on therapists is needed because between-therapist differences seem to be associated with patient outcomes over differences within therapists. While the study confirmed that good early working alliance improves outcome during treatment without being linked to post-treatment recovery, more research is needed also in that area. (Contains 6 tables and 1 figure.)