Corporate data of poor quality can have a negative impact on the performance of business processes and thereby the success of companies. In order to be able to work with data of good quality, data quality requirements must clearly be defined. In doing so, one has to take into account that both the provision of high-quality data and the damage caused by low-quality data brings about considerable costs. As each company’s database is a dynamic system, the paper proposes a cybernetic view on data quality management (DQM). First, the principles of a closed-loop control system are transferred to the field of DQM. After that a meta-model is developed that accounts for the central relations between data quality, business process perfor-mance, and related costs. The meta-model then constitutes the basis of a simulation technique which aims at the explication of assumptions (e.g. on the effect of improving a data architecture) and the support of DQM decision processes.