Ambulance services produce a large quantity of data, which can yield valuable summary statistics. For strategic planning purposes, an economic framework is proposed, and the following four resource allocation questions are answered, using data from the Surrey Ambulance Service: (1) To satisfy government response time targets, how many additional ambulances will be required, ceteris paribus? (2) To minimize average response time (r*) with given resources, how should ambulances be rostered temporally? (3) Which innovations are worth undertaking? (4) How would an increase in demand affect r*?The 'Ambulance Response Curve' --the relation between response time and the number of available but not-in-use ambulances--is used to estimate how much r* will be reduced by deploying an additional ambulance. Estimating the marginal cost of an ambulance allows us to estimate the opportunity cost of each second of response time, and to compare the cost of three 'innovations' with that of increasing resources. The time savings of adding an extra ambulance at each of the 168 h of the week are examined.In 1997-1998, r* was 8 min 52 s. An additional ambulance reduces r* by 8.9 s. Each reduction of 1 s in r* costs 28,000 pounds per year. Fourteen additional ambulances are required to meet response time targets if the 8.9 s reduction per ambulance is maintained. r* reduces by 4.6 s when ambulances are shifted from early mornings to Saturday evenings. Activation time reduces by 38 s when crews sit in their ambulances. A 1 min decrease in overall call time decreases r* by 1.1 s. Answering only 10 per cent of all calls reduces r* by 63 s. An increase of demand of 10 per cent increases r* by 7.8 s.Ambulance services will be better able to determine which innovations are worth undertaking. Policy makers will be better placed to determine funding levels to achieve response time targets.