Metadata management is a foundational component of a successful data strategy. Metadata supports data governance, regulatory compliance and data management demands by providing critical details about data assets such as: what it is, when it was created, how it’s been altered, and who has access. While metadata may sound like technical jargon, in simple terms, metadata is used to “label” data with information to organize, classify and, ultimately, to understand data.
Without metadata, businesses cannot manage the massive amounts of diverse data collected across their enterprise. Metadata is critical for understanding and effectively deploying data resources to support enterprise departments and their processes. Furthermore, metadata provides the essential information required to enable advanced analytics.
The most tangible output of metadata management is a metadata glossary that lists the location of an organization’s data as well as essential information on how to use it. Before building this dynamic data “knowledge base”, organizations must systematize and interpret metadata from various perspectives.
To achieve a comprehensive understanding of where data lives in the business, and how it’s deployed, organizations must gather and arrange metadata from the physical, logical and conceptual perspectives or ”lenses.” Metadata from the physical perspective (or, physical metadata) details the specifics of where data sets live, such as the system it resides in, down to the schema, table and column, or key-value level of detail. This type of metadata is typically machine generated and can be automatically grabbed from software and systems.
Logical metadata provides details regarding how data is woven together to form larger sets and how data flows through systems and processes, from intake or creation, through storage and transformation, and finally, on to consumption. It establishes a roadmap on data’s path through the data supply chain, including its usage and transformation over time.
Conceptual metadata provides the business context for data, detailing its meaning and purpose within the organization. Conceptual metadata provides critical information about the usage of data across the enterprise, and represents the accumulated knowledge of subject matter experts within the organization. Because it is derived from people, it is the most difficult type of metadata to collect and update, as it requires human labor and management processes to continuously update, though some new technologies can assist this process by leveraging machine learning algorithms to populate some metadata tags and to ensure overall reliability.
By viewing data with these three lenses–physical, logical and conceptual–the metadata framework starts to form. The next step is to ensure the data governance model empowers the business to organize the metadata glossary and make it readily available to all users.
To create a comprehensive metadata glossary that gives both business and technical users transparency into their data, organizations need to invest in a data governance strategy. The strategy must promote open communication among data owners, data stewards and data consumers to foster a community approach to data meaning and comprehension. When different teams join forces to define and document metadata, it builds a shared understanding of data assets and it reduces the confusion that business users may face when searching the metadata glossary for the first time.
Data governance starts by assigning accountability for individual data assets. It must establish clear lines of responsibility to ensure that metadata stays accurate and that the data is accessible to authorized users. With full participation across the enterprise, data governance delivers complete transparency into an organization’s data landscape. Business users are empowered to easily and quickly define, track and manage all aspects of their data assets.
Data Governance practices must evolve to keep pace with the ever-growing supply of and demand for data. For example, data governance must support advanced analytics, such as machine learning, that can consume huge quantities of data, by helping to supply reliable data enriched with comprehensive metadata. Fortunately, you can leverage some of the same new technologies to aid in governance by providing surveillance over data and by automating the capture and curation of some metadata.
With the right data governance strategy, enterprises can successfully create a comprehensive metadata glossary which will empower all users across the business to do more with data.
Are you looking for additional information about organizing, defining and understanding metadata? Check out the e-book below.
For a deeper dive into this topic, visit our resource center. Here you will find a broad selection of content that represents the compiled wisdom, experience, and advice of our seasoned data experts and thought leaders.