summary Two groups independently carried out qualitative studies of the development, implementation and evaluation of telehealth systems and services in the UK. The data collected (in more than 600 discrete data collection episodes) included semistructured interviews, observations and documents. We conducted a conjoint reanalysis of the data. The objective was to identify the conditions which dispose a telehealth service to be successful or to fail. There appear to be four conditions necessary for a telemedicine system to stabilize and then normalize as a means of service delivery. When one or more is absent, failure can be expected. These conditions are often overlooked by local proponents of telemedicine, who seem to rely on demonstrations that the equipment works as the primary criterion of success.