No abstract

No abstract

Data consumers assess quality within specific business contexts or decision tasks. The same data resource may have an acceptable level of quality for some contexts but this quality may be unacceptable for other contexts. However, existing data quality metrics are mostly derived impartially, disconnected from the specific contextual characteristics. This study argues for the need to revise data quality metrics and measurement techniques to incorporate and better reflect contextual assessment. It contributes to that end by developing new metrics for assessing data quality along commonly used dimensions - completeness, validity, accuracy, and currency. The metrics are driven by data utility, a conceptual measure of the business value that is associated with the data within a specific usage context. The suggested data quality measurement framework uses utility as a scaling factor for calculating quality measurements at different levels of data hierarchy. Examples are used to demonstrate the use of utility-driven assessment in real-world data management scenarios and the broader implications for data management are discussed

No abstract

A large majority of work in database marketing deals with what to do with data when it is available. This paper focuses on an aspect of data that has not been visited frequently in the database/interactive marketing literature—managing the quality of data resources from a profit point of view. While costly to achieve and sustain, high data quality is essential for effective database marketing. The notion that “more is better” very often prevails in data quality management decisions, essentially with very little consideration, if any, of cost. This paper suggests that data quality management decisions should be driven by considerations of cost–benefit tradeoffs and profit maximization. It specifically addresses data quality decisions which are highly relevant in the database marketing area: the subset of data records managed, reflecting time-span coverage, and the targeted quality levels in this subset. Decisions of these types are routinely made based on satisfying technical and functional requirements. In this study, we propose a model that quantifies the benefits and the costs associated with these decisions, and permits optimizing them from a profit maximization standpoint. The paper describes the model development, discusses its implications for data quality management decisions, and highlights its potential contributions with illustrative examples.