In numerous client engagements, we have witnessed that Data Quality (DQ) is a never-ending battle. At many companies, IT fixes and re-fixes the data rather than develop solutions and manage applications. DQ is not confined to IT, but maintaining that quality requires effort from all data users, especially those at the business end.
Building a company-wide initiative can be a hard sell, as the enormity and the scope are complex. However, applying some proven strategies helps create awareness and gains support for DQ. The key idea is to avoid fixing the problem over and over and instead understand and communicate the bigger picture to solve the problem as IT and Business Information management mature.
Setting up key quality measures and documenting impacts are the best ways to gain support. As part of any initiative, include DQ measures that can be gathered and reported periodically. Having information metrics in hand is sure way of getting attention. Recommended measures include:
- Down time caused by quality issues
- Man hours invested in repeat problems
- Ownership of the quality issues
Any data project should consider the trust aspects of the data. Introducing certification processes for adding new data, especially large-batch data, tremendously improves business participation and overall quality. Errors such as missing/null values and wrong information (invalid values), rejections, and warnings should be communicated and remedied within expected time frames. Often the fixes belong in the up-stream systems, which eliminates current and future data issues.
Creating a data certification process engages the business and gains trust. The important point here is making business, not IT, responsible for the data.
DQ as part of SDLC
Data quality should be part of the software development life cycle to ensure acceptable quality. Development projects should incorporate time for reporting on quality measures. Doing this improves data trustworthiness and speeds adaptation of new applications.
Governance is the best way to gain support for quality initiatives. Many times, when cost-cutting measures are enacted, quality control during development suffers because it is viewed as a nice procedure to have but not essential. Through data governance, mandated quality control measures can gain necessary support. So, it is important to have a process in place that ensures data governance and leverages it to bring about data quality transformation.
Incremental changes that emphasize quality in existing processes help build a case for broader process changes that are mindful of quality.