A Mathematical Approach to Data Quality Management in Enterprise Systems

The quality of data contained in enterprise information systems has significant impact on both internal business decision-making and external regulatory compliance. This paper addresses the issue of data quality management in information systems within an enterprise. Motivated by such legislative mandates as the Sarbanes-Oxley Act of 2002 on the reliability and integrity of the information system and the data produced by these systems, we propose a process-based modeling framework to assess the impact of error propagation and the resulting data quality metrics. This framework is then integrated with a business controls framework in which the placement and effectiveness of control procedures alter the propagation of errors and ultimately the quality of the data in the business process. This integrated modeling framework enables mathematical formulations of managerial problems that lead to effective data quality control strategies. We develop a two-stage multiple-choice knapsack model as a special case, and illustrate the model and analysis with a real-world procurement process