Rehabilitation research is challenged to improve its evidence base, which requires more precise and more consistent conceptualisation and measurement of treatment ingredients. This paper presents the steps in defining and specifying treatments towards the construction of a therapy manual for experimental interventions, or a coding system for observational treatment research. Specifying the active ingredients, identifying the behavioural operations that are aligned with each, and developing procedural details and materials are discussed and illustrated using two on-going trials for treatment of traumatic brain injury. The process of assessing treatment fidelity, that is, the extent to which interventions are delivered, received and used as intended, is also discussed.