William wishes to thank Stuart Mullins, Lucidity Consulting Group, for his contribution to this month's column.

In my April column, I discussed the problems that organizations face in defining data quality and classifying what kind of problem exists. I listed seven sources of data quality failures and discussed the first three, which included entry quality, process quality and identification quality. I'll continue the discussion around the four remaining sources of data quality issues and discuss the components required to develop a plan of action. 

Integration Quality 

Integration quality, or quality of completeness, can present big challenges for large organizations. Integration quality problems occur because information is isolated by system or departmental boundaries. It might be important for an auto claims adjuster to know that a customer is also a high-value life insurance customer, but if the auto and life insurance systems are not integrated, that information will not be available. 

While the desire to have integrated information may seem obvious, the reality is that it is not always apparent. Business users who are accustomed to working with one set of data may not be aware that other data exists or may not understand its value. Data governance programs that document and promote enterprise data can facilitate the development of data warehousing and master data management systems to address integration issues.  

MDM enables the process of identifying records from multiple systems that refer to the same entity. The records are then consolidated into a single master record. The data warehouse allows the transactional details related to that entity to be consolidated so that its behaviors and relationships across systems can be assessed and analyzed. 

Usage Quality

Usage quality often presents itself when data warehouse developers lack access to legacy source documentation or subject matter experts. Without adequate guidance, they are left to guess the meaning and use of certain data elements. Another scenario occurs in organizations where users are given the tools to write their own queries or create their own reports. Incorrect usage may be difficult to detect and quantify in cost.  

Thorough documentation, robust metadata and user training are helpful and should be built into any new initiative, but gaining support for a post-implementation metadata project can be difficult. Again, this is where a data governance program should be established and a grassroots effort made to identify and document corporate systems and data definitions. This metadata can be injected into systems and processes as it becomes part of the culture to do so. This may be more effective and realistic than a big-bang approach to metadata.

Aging Quality 

The most challenging aspect of aging quality is determining at which point the information is no longer valid. Usually, such decisions are somewhat arbitrary and vary by usage. For example, maintaining a former customer's address for more than five years is probably not useful. If customers haven't been heard from in several years despite marketing efforts, how can we be certain they still live at the same address? At the same time, maintaining customer address information for a homeowner's insurance claim may be necessary and even required by law. Such decisions need to be made by the business owners and the rules should be architected into the solution. Many MDM tools provide a platform for implementing survivorship and aging rules.

Organizational Quality 

Organizational quality, like entry quality, is easy to diagnose and sometimes very difficult to address. It shares much in common with process quality and integration quality but is less a technical problem than a systematic one that occurs in large organizations. Organizational issues occur when, for example, marketing tries to "tie" their calculations to finance. Financial reporting systems generally take an account view of information, which may be very different than how the company markets the product or tracks its customers. These business rules may be buried in many layers of code throughout multiple systems. However, the biggest challenge to reconciliation is getting the various departments to agree that their A equals the other's B equals the other's C plus D.

A Strategic Approach

The first step to developing a data strategy is to identify where quality problems exist. These issues are not always apparent, and it is important to develop methods for detection. A thorough approach requires inventorying the system, documenting the business and technical rules that affect data quality, and conducting data profiling and scoring activities that give us insight in the extent of the issues.  

After identifying the problem, it is important to assess the business impact and cost to the organization. The downstream effects are not always easy to quantify, especially when it is difficult to detect an issue in the first place. In addition, the cost associated with a particular issue may be small at a departmental level but much greater when viewed across the entire enterprise. The business impact will drive business involvement and investment in the effort.

Finally, once we understand the issues and their impact on the organization, we can develop a plan of action. Data quality programs are multifaceted. A single tool or project is not the answer. Addressing data quality requires changes in the way we conduct our business and in our technology framework. It requires organizational commitment and long-term vision. 

The strategy for addressing data quality issues requires a blend of analysis, technology and business involvement. When viewed from this perspective, an MDM program is an effective approach. MDM provides the framework for identifying quality problems, cleaning the data and synchronizing it between systems. However, MDM by itself won't resolve all data quality issues. 

An active data governance program empowered by chief executives is essential to making the organizational changes necessary to achieve success. The data governance council should set the standards for quality and ensure that the right systems are in place for measurement. In addition, the company should establish incentives for both users and system developers to maintain the standards.  

The end result is an organization where attention to quality and excellence permeate the company. Such an approach to enterprise information quality takes dedication and requires a shift in the organization's mindset. However, the results are both achievable and profitable.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access