13.2 Moving to a Position of High Data Quality Requires an Explicit Effort

Note

Data quality deserves an explicit quality assurance function that is fully supported by management. It is an ongoing requirement.

Getting to a position of high data quality and maintaining it are complex tasks that require dedicated professionals who are armed with appropriate methodologies and tools. It takes a lot of work to fix quality problems and to prevent them from recurring in new ways.

Note

Data quality must become a part of everyone's job. The education and inspiration must come from the data quality assurance group.

The data quality assurance group must have the cooperation and effort of many different people to be effective. They need business analysts, subject matter experts, data architects, database administrators, data entry personnel, and others to cooperate and contribute. If they operate entirely within themselves, they cannot succeed.

Note

The most effective way to organize a data quality assurance function is to work with the data first to find inaccuracies, research them to create issues, and then monitor the progress of issue resolution.

Many data quality issues lurk in the data without obvious external manifestations. These can often be dug out and used to build a case for making system improvements. Even when issues come from the outside, they need to be investigated in the data to find the extent of the problem and to find related problems that are not as obvious. The data will tell you a lot about itself if you only listen.

Note

A large return on investment can be realized through helping new initiatives improve the quality of the data they are working with and avoid making new data quality problems through inappropriate transformation or use of data.

Pure data quality assessment projects have difficulty in gaining approval because of the low promise of return and the disruption it brings to operational environments. Having a data quality assurance team work with funded projects that are trying to change or extend a system can generally return more value to the project than the cost of assessment. In addition, they can improve the quality of the data at the same time. This is a win-win strategy for everyone.