“Metadata is our friend,” said Hitachi’s CTO at Meritalk’s Big Data Brainstorm late last year.
He was talking about how metadata can be used to create a tiered storage system to prevent data lakes from turning into data swamps. But he’s also right in that metadata should be treated like a friend. If you take it for granted, it’s probably not going to be your friend for much longer.
I would venture to guess that most organizations aren’t exactly in the best terms with their metadata. They’ve been neglecting it far too long.
Metadata, of course, are tags that describe the data those tags are attached to. With proper metadata management, organizations can essentially sort through and group their data by all sorts of definitions and categories to make finding and using certain data assets that much easier. With proper metadata management, you can prevent data access or retrieval issues, you can clamp down on redundancy, you can avoid poor governance issues, and you don’t have to worry about tapping particular people (who may not remember or have left the company) to explain the origin or value of certain data. But most of all, you can ensure your data can be easily found and used at the moment it is most valuable.
But you can’t treat your data like a fixed commodity. You don’t eat stale bread, do you? Then why do you support stale metadata? You can’t wait months or years (like most organizations do) to update your metadata. It must be reviewed and updated constantly to keep it from becoming stale or sinking once again into chaos. Data changes and so does the metadata associated with it. Similarly, so do the regulations surrounding the protection of certain data changes. Failing to adhere to new rules is bad data governance, which can in turn open your business up to a variety of risks, including massive fines, legal liabilities and hits to the company’s reputation or brand.
But there’s something else: even if the data doesn’t change, certain external forces may cause its value to change. What data may mean to the business one day may change the next. Failing to stay on top of this means letting high value data go unrecognized and unused, which could lead to major setbacks to the business—competitive and otherwise—in today’s real-time data-driven world.
Maximizing the value and use of your data starts with proper metadata management. This means:
1. You should ensure metadata is generated as data is created, collected or changed. Incorporating the use of a smart data catalog, which can save time by quickly tagging data and updating those tags even as data comes in through a combination of crowdsourcing and machine learning, helps address this point.
2. Your data analysts should have a metadata system that adheres to best practices for creating and tracking metadata.
3. You should constantly review metadata to ensure it remains both current and relevant.
Remember, strong data management comes down to strong metadata management. Structuring your data management and keeping it structured is key to transforming your business into a powerful data-driven enterprise. If you want to stay friends with your metadata, treat your metadata right, and it will be a good friend to your organization for a long time to come.