In cellular communication networks, freedom to roam is a fundamental quality of service requirement. This is governed by the spatial availability of adequate received signal strength. However, base station procurement and operation is a significant cost for the operator and deciding on the most appropriate level of coverage is problematic. We introduce a computational framework to address the marginal cost of service coverage. This can be applied to particular real world planning scenarios to assess the rate of change of cost with respect to service coverage. Analysis of this function determines coverage levels beyond which increased investment yields minimal additional spatial service provision. We present a case study of two benchmark real-world planning scenarios for GSM networks. The model we apply incorporates network traffic, propagation-loss data from the eld and constraints concerning inter-cell interference on subscribers. We explain how the framework is of practical use for the operator during rollout.