An important problem with the use of OR-solutions for repetitive decision making is how to control the quality of the solution, if the problem to be solved may change with time. Current methods are generally lacking procedures for diagnosing the causes of changes in the quality. In this paper we use concepts borrowed from general control theory to derive a scheme for monitoring the solution. This relates changes in the performance to changes in the quality of the model used, of the decision routine used, of the information used, and in the usage of the solution. It is shown that the existence of a model of the process to be controlled (a design process model), is crucial for designing adequate monitoring procedures. An example of the application of the monitoring scheme to a production control problem in practice is presented and the experience obtained from its use are discussed.