Background Social anxiety disorder (SAD) is common, typically starts in adolescence and has a low natural recovery rate. Existing psychological treatments for adolescent SAD are only moderately effective. It is possible that recovery rates for adolescents could be substantially improved by adapting a psychological therapy that is highly effective among adults with SAD. Objectives To train child and adolescent mental health services (CAMHS) therapists to deliver cognitive therapy for SAD in adolescents (CT-SAD-A) and assess therapist competence. To estimate the costs to the NHS of training therapists to deliver CT-SAD-A and the mean cost per adolescent treated. To examine the feasibility of a randomised controlled trial (RCT) to compare CT-SAD-A with the general form of cognitive–behavioural therapy that is more commonly used. Design During the training phase of the study, it became clear that the RCT would not be feasible because of high staff turnover and unfilled posts within CAMHS and changes in the nature of referrals, which meant that few young people with primary SAD were accessing some of the participating services. The study design was altered to comprise the following: a training case series of CT-SAD-A delivered in routine CAMHS, an estimate of the cost to the NHS of training therapists to deliver CT-SAD-A and of the mean cost per adolescent treated, and qualitative interviews with participating young people, parents, therapists and service managers/leads. Setting Five CAMHS teams within Berkshire Healthcare and Oxford Health NHS Foundation Trusts. Participants Eight therapists received training in CT-SAD-A. Twelve young people received CT-SAD-A, delivered by six therapists. Six young people, six parents, seven therapists and three managers participated in qualitative interviews. Interventions Cognitive therapy for social anxiety disorder in adolescents (CT-SAD-A). Main outcome measures Measured outcomes included social anxiety symptoms and diagnostic status, comorbid symptoms of anxiety and depression, social and general functioning, concentration in class and treatment acceptability. Patient level utilisation of the intervention was collected using clinicians’ logs. Results Nine out of 12 participants achieved good outcomes across measures (r ≥ 0.60 across social anxiety measures). The estimated cost of delivering CT-SAD-A was £1861 (standard deviation £358) per person. Qualitative interviews indicated that the treatment was acceptable to young people, parents and therapists, but therapists and managers experienced challenges when implementing the training and treatment within the current CAMHS context. Limitations Findings were based on a small, homogeneous sample and there was no comparison arm. Conclusions CT-SAD-A is a promising treatment for young people with SAD, but the current CAMHS context presents challenges for its implementation. Future work Further work is needed to ensure that CAMHS can incorporate and test CT-SAD-A. Alternatively, CT-SAD-A should be delivered and tested in other settings that are better configured to treat young people whose lives are held back by SAD. The new schools Mental Health Support Teams envisaged in the 2017 Children’s Mental Health Green Paper may provide such an opportunity. Funding The National Institute for Health Research (NIHR) Health Technology Assessment programme. Individual funding was also provided for Cathy Creswell, David M Clark and Eleanor Leigh as follows: NIHR Research Professorship (Cathy Creswell); Wellcome Senior Investigator Award (Anke Ehlers and David M Clark); and the Wellcome Clinical Research Training Fellowship (Eleanor Leigh).