“You keep using that word. I do not think it means what you think it means.”
That quote by Inigo Montoya (portrayed by Mandy Patinkin in The Princess Bride) has passed through every data manager’s mind since 1987. Even if you haven’t seen the movie, you get the idea about what’s being said: Two people have (very) different ideas about what some word or phrase means.
In data management circles, we use a Business Glossary to align everyone to a single meaning of a business term or data element so that reporting and analytics are consistent across departments and business units.
A “customer” is _______________.
A “product” is _____________.
“Risk rating” means ______________.
“Past Due Account” means _____________.
But what happens when an organization changes the meaning of a data element? Business moves forward and requirements evolve. So should the definitions in our Business Glossary – that’s the point, and nobody disputes that. But one of the things I see organizations struggling with is how to maintain a history of those changes. We always identify the business definition (along with technical and process-related definitions) as a critical component of our metadata, but maintaining historical metadata over time is something that gets overlooked. When selecting your metadata management tools, versioning is an important feature to consider.
Maintaining a record of what changed is a key step of any data governance or data management process. Data consumers need to know what any data element means, but they also need to know what it meant in the past and why it might have changed. Anyone comparing historical data to the current period will need to know that. Auditors and regulators care deeply about how definitions and business rules change over time. Maintaining that history and making it easily accessible to all data stakeholders keeps everyone aligned to those changes.
Versioning is perhaps one of the strongest justifications to maintain your metadata in something besides spreadsheets. True, SharePoint and other document management systems can maintain versions of a spreadsheet, but the user experience is less than optimal. Finding the right historical spreadsheet can be a challenge, and identifying exactly what the change was can be impossible.
Besides presenting the current view of your metadata, the ideal metadata platform will: (a) show the historical metadata for any given element; and (b) pro-actively notify stakeholders of changes when they occur. Your metadata platform has to be a time machine that can take users to any point in the past with a few mouse clicks. The more easily these systems handle versioning and notifications, the easier it will be to keep everyone aligned to changes. Change happens. Plan for it. Embrace it.