Best practices: Business value, business rules and information quality metrics
Presuming that we all agree that poor information quality can affect the bottom line, the more pertinent questions are to understand how and how much poor information quality costs the organization. When talking to representatives of a company, it is easy (and perhaps simplistic), to claim that because their company is not unlike others within the same industry, the costs associated with their poor information quality are going to be similar. However, I have found in my conversations with senior managers that there is always a perception that their company is not like other companies, and consequently, their problems are unlike any other organization in the world.
To address this perception, it is desirable to evolve a set of information quality metrics that directly correspond to ways the organization is not meeting its business objectives. Unfortunately, most approaches of measuring information quality are tied to ways that vendors compare their products (either relating to counting duplicate records, recording the number of non-USPS compliant addresses, or basic frequency analysis of data values), which is deficient for two reasons. First, invalid addresses or missing values are likely evidence of information quality problems, but they do not always represent the problems themselves. Second, these measurements are only half of what a reasonable metric would provide: they give you a quantification of the evidence of the problem, but do not convey any indication of the reduction in performance attributable to that problem.