There is a global search for performance indicators for health services as a basis for quality improvement, external assessment, management control and public accountability and information. This requires agreement on what constitutes good performance as perceived by a disparate range of potential users, and on what aggregated or tracer data faithfully reflect policy objectives. To these challenges are added the technical problems of collating comparisons which are complete, accurate, timely and statistically valid, as well as behavioural problems of their interpretation and use. Examples are given of the Patients' Charter (National Health Service in England) and the Scottish outcome indicators. This experience from the UK may well have some relevance in other countries.