Despite advances in the development of voice outcome measures, there is no methodology to define and quantify the elements of the complex process of speech therapy. The components of therapy given by one therapist to the intervention limb of a controlled trial werecharacterized according to a list of five minutes. The intervention was of proven benefit compared with a control period of observation. Indirect approaches comprised two thirds of therapy time. The types and duration of intervention were assessed but no treatment category seemed moreassociated with a favourable outcome. The design shows that it is possible to perform a prospective, structured analysis of the components of voice therapy. The method appears viable for thefuture comparison of the widely varying techniques current in voice therapy practice.