Data Alignment and Fitness Assessment

It is widely accepted that data quality is about fitness for use. Therefore, the data quality baseline needs to be analyzed with that perspective in mind. It is not about achieving the score of 100 for every single element in the baseline. It is about achieving the proper score necessary to keep the business operating efficiently and confidently. Moreover, it is about identifying areas for data quality improvements.

Even though the initial assessment has a potentially non–business oriented component attained by the bottom-up approach, this step takes care of aligning the results to a specific business purpose. Simply speaking, the baseline is a superset, which can now be properly aligned.

It is expected the business will look at some of the results and say: “We had no idea!” Another expected outcome is the need to look further into a given element. For example, when the business finds out only 60 percent of the U.S. postal codes follows the proper pattern of five digits, a dash, plus four digits, they might want to know what other patterns exist.

Several interactions to create, analyze, and refine the baseline will most likely be required. But a fully finalized baseline is not necessarily required before actions can be taken. The baseline will have areas more clearly defined than others. If those areas require corrections, the proper data quality initiative should start as soon as possible. It obviously needs to be properly prioritized according to other ongoing activities and proper resource availability, but they don't necessarily all need to be part of a single data quality project. As a matter of fact, data quality projects can be very resource intensive and will need to be prioritized properly in consideration of other organizational activities.

It is important to decide what problem to tackle and how much of the data needs to be fixed. Use commonly known tactics such as low-hanging fruit and the Pareto principle, aka the 80-20 rule, when prioritizing tasks and deciding what, when, and how much to fix. Data issues will always exist; consequently, focus the effort where it is most needed and where there will be the most impact on the business.