A procedure for the optimal determination of measured variables in chemical processes is presented. This procedure is based on quantifying the trade-off between the process information that can be obtained and the total cost of the sensors making up the measurement system. The resulting trade-off is specified as a Pareto optimization problem. The optimal measurement systems that should be considered for installation are determined from the set of solutions to this Pareto optimization problem. The approach is demonstrated using a CSTR process example.