Objectives. To examine the clinical effectiveness of guided self-help (GSH) for anxiety and depression in routine clinical practice, and the role of self-efficacy, therapeutic alliance and socio-economic status in influencing that effectiveness. Design. A within-subjects repeated measures design in which participants served as their own controls by completing questionnaires across a control period prior to GSH intervention, then again at post-intervention and 3- and 6-month follow-up. Methods. GSH participants completed outcome measures for mental health (HADS) and work/social functioning (WSAS). Factors explored by regression as possible predictors of effectiveness were self-efficacy, therapeutic alliance and socioeconomic status. Results. Sixty people completed GSH, with analyses indicating effectiveness of GSH in significantly improving mental health and social functioning at post-treatment and 3-month follow-up, but not at 6-month follow-up. Effectiveness was also indicated under intent-to-treat conditions (n = 97) with medium effect sizes (≈ 0.6) for each outcome measure at post-treatment. Improvement in mental health was predicted by lower self-efficacy and greater therapeutic alliance. Completers of the intervention had significantly higher socio-economic status than non-completers. Conclusions. The current study has suggested effectiveness of GSH in routine clinical practice across different primary care services at post-treatment, but with less evidence of this at follow-up. Effectiveness has been highlighted to be influenced by self-efficacy and therapeutic alliance, suggesting the importance of considering non-specific factors when patients access GSH in primary care. This study underlines the need for further research exploring longer-term clinical effectiveness and examining for whom GSH works in order to constructively inform future evidence-based practice.