13.3 Data Accuracy Is the Cornerstone for Data Quality Assurance

13.3 Data Accuracy Is the Cornerstone for Data Quality Assurance

Although data quality has many dimensions, data accuracy is the foundation dimension. If the data does not represent the true facts, all other dimensions are less important.

Using the inside-out approach by performing extensive analysis over the data is a fast and efficient way to dig out a lot of data inaccuracy issues. It cannot find all issues because it is always possible that data satisfies all rules and is still wrong. However, the majority of bad practices in entering data result in telltale conditions in the data that can be exposed through application of rules.

In addition to finding a lot of inaccurate data, the data profiling process described in this book will also produce an accurate metadata description of the data. This description will generally be more complete and more accurate than any descriptions previously held.

The information produced through data profiling technology often provides the clues to the right solutions for improving and monitoring systems.

This is not a definitive book on data quality. It focuses on an aspect that has not been given much coverage but promises to provide major value to corporations. The focus on data accuracy and data profiling technology as a means of improving it is a powerful message to those wanting to do something about poor data quality.

The time has come for corporations to move their interest in the quality of data to a new level. You can expect most corporations to invest in new initiatives aimed at this topic. A failure to do so can mean your competitors will leave you behind.

The big gains are not in the small improvements to your environment, wherein you eliminate shipping the wrong product occasionally, avoid losing a customer from time to time, or get too low a discount from a supplier. The big gains come from being able to respond to business opportunities faster with systems that are resilient and flexible. To be flexible, your systems have to be well documented, and well understood, and the data must be as squeaky clean as possible.

The corporation that implements a new technology or business model in one year will be better off than one that takes three to four years. A corporation that can consolidate data from a merger or acquisition in nine months will beat the competitor who takes three years.

You can achieve high performance in implementing new business models if you make significant improvements in the accuracy of your data and in the metadata that supports it. If nothing is taken from this book other than one fact, it should be this: