Data accuracy is the most visible and dramatic dimension of data quality. It is the easiest to expose, the easiest to make improvements in, often does not require system reengineering to achieve improvements, and often does not require reorganization of your corporation to accommodate it. Although you cannot get to perfect accuracy in your data, you can improve the accuracy to the point where it consistently provides information that drives correct decisions.
Data accuracy is a complex subject that needs to be fully understood. The concepts of valid versus invalid, inconsistencies in representation, object-level inconsistency, representation of values not known, and missing information are all part of defining accuracy.
There are two methods of determining the accuracy of data: reverification and data analysis. Neither one can guarantee finding all inaccurate values.
Reverification is generally too expensive and slow to be effective. Analytical techniques are easier to use. Analytical techniques require that you understand what the definition of "correct" is for each data element. Just as accurate data is the foundation component of data quality, analytical techniques are the foundation component of any effective data quality assurance program.