Our Loraine Lawson continues her discussion of data quality and controls with the news that some middleware experts estimate the U.S. health care industry loses $314 billion annually to fraudulent or inaccurate billing. Loraine goes on to describe some tips on cleaning up data during integration projects, and then asking why issues like data deduplication are not higher up on many industries' priority list.
Here in the IT Downloads library, we have two resources that will help you evaluate the health of your own data and begin the process of establishing a culture of data quality assurance, which ultimately is the first (and potentially most difficult) step in the process. Both these resources are available free to IT Business Edge members.
"Nuts & Bolts: The Mechanics of Data Governance," an e-book from our partners at Baseline Consulting, begins its discussion with the importance of getting executive sponsorship and buy-in for your data quality initiative. Any meaningful data governance effort will change the way some folks do their jobs; ideas for how to best accomplish this may come from the bottom up, but the mandate needs to come from the top down.
The e-book goes on to describe the differences between data stewardship, which is an oversight function, and data management, which Baseline Consulting breaks into six primary operational areas, which you can see in the figure below.
The e-book then goes on to describe each phase of data management in detail. For data quality, Baseline emphasizes the importance of ongoing monitoring and metric reporting, which sometimes can get lost in the shuffle of all the integration projects most big shops are juggling. Data stewards often are charged with data quality monitoring, but other staff members, including your business analysts, should be in the loop to create a system of checks and balances.
The Data Quality Health Check Tool, from our partners at Info~Tech Research Group, asks you a series of questions about the current state of data quality in your business. Some example questions are:
Obviously, users can cause a lot of problems by incorrect data input. But a well-designed input system can mitigate a lot of that grief. For example, you can scan the notes field for critical keywords and pop up a warning message suggesting to users that they include that info in a structured field.
Based on your answers to the 15-question survey, the Excel-based tool gives you a ranking of your data's health and suggests certain actions. If things are bleak enough, it advises you to designate someone as a data steward; if your enterprise is really serious about data quality, you may want to just go ahead and create that position, before serious problems arise.