The organisation, premises, and staffing of three child health clinics of different types in an inner city area have been studied and related to information on immunisation derived from their computerised child register and other returns, together with some simple estimates of cost. Differences were found between overall uptake of different antigens, consent rate, compliance after consent had been given, and the median age at immunisation for children allocated to each clinic. Thus consent to pertussis immunisation varied from 90% in one clinic to 69% in the other two. In the clinic with the high consent rate, however, only 66% of those whose parents had consented actually received all three doses compared with 82% and 85% in the other clinics. The median age of immunisation also showed pronounced differences between the clinics--314, 351, and 375 days, respectively, for the third dose of the triple course. There were differences in characteristics of the clientele served by each clinic and in staffing and facilities, which led to differences in costs, as well as contributing towards the variable patterns of uptake. We suggest that indicators such as these, largely available from the child health computer system, are used systematically to identify ways of improving uptake and increasing cost effectiveness.