Metadata: what it is, the types, and why it underpins data governance

Everyone has used metadata without realizing it. When you look at a photo on your phone and see the date, the location, and the camera model, you're reading metadata. When you open a spreadsheet and the column says "registration date," "owner," and "last updated," that's metadata too. Metadata is the data that describes other data. It sounds like a technical detail, but inside a company it decides whether people can find, understand, and trust information. And that is exactly where most organizations stumble.

What metadata is

Metadata is information that describes, explains, or gives context to a piece of data. If the data is the number "1,500,000" in a cell, the metadata is what answers the obvious questions: is this revenue or cost? Is it in reais or dollars? Which month does it refer to? Who generated this number, and when? Without metadata, a data point is just a loose value, and a loose value doesn't support any decision.

In business practice, data is the content and metadata is the label. A material master has the code, the description, and the unit of measure as data. What that description means, which area owns that record, when it was created, what the data-entry rule is, and where it came from: all of that is metadata. It's the difference between having a database full of fields and having a database that people actually understand.

The types of metadata in practice

People who work with governance usually split metadata into three big groups, and it's worth understanding each one because they solve different problems.

Business metadata translates the data into the language of whoever uses it. It's the definition of "active customer," the explanation of what goes into the calculation of "net revenue," the glossary that says what the company considers an "approved supplier." This is the type that companies lack the most and the one that creates the most confusion, because each area ends up inventing its own definition.

Technical metadata describes the structure of the data in the systems. Table name, field type, size, format, key, relationships between tables. It's what IT uses to integrate systems, migrate data, and keep everything running underneath.

Operational metadata tells the story of the data over time. When it was created, who changed it, when the last load was, how many records came in, whether there was a processing error. It's the type that lets you audit and trust, because it shows whether the data is alive or has been sitting idle for three years without anyone noticing.

There is also a layer that stitches all of this together, which is data lineage, that is, the path the data travels from its origin to the final report. Knowing that a number on the director's dashboard came from a record entered by hand by an intern completely changes the level of trust you can place in that number.

How metadata works day to day

The value of metadata shows up in the most mundane moment of work, which is when someone needs a piece of information and doesn't know where it is. A financial analyst looks for the margin report and finds five different versions, each with a different total. Which one is right? Without metadata, he'll ask around the hallway, message the group chat, and in the end pick the one that looks most reasonable. With well-defined metadata, he sees which is the official source, who is responsible, when it was updated, and how the number is calculated. The decision stops being a guess.

This happens a lot in SAP projects. When the company is about to migrate to S/4HANA, someone needs to know what each master data field means, what the data-entry rule is, and which data can move over and which must be fixed first. If that knowledge lives only in the heads of two or three people, the project becomes hostage to them. Documented metadata takes that knowledge out of people and puts it into the organization.

Who takes care of metadata

Metadata is not just an IT matter, and that's a common mix-up. The technology area takes care of the structural side, the technical metadata, the infrastructure that stores and connects everything. But the one who defines the meaning of the data is the business area. The purchasing team knows what a critical supplier is. Finance defines recognized revenue. Sales says what counts as an active customer.

In the governance structure, the data owner is the owner of the data in the business area and answers for the definitions. The data steward keeps all of this organized day to day, makes sure the glossary is up to date, that fields are documented, and that rules are followed. And the data or governance area provides the platform and the method so that this knowledge isn't lost. Without these roles defined, the metadata may exist, but it's scattered, outdated, and ownerless.

Which tools help manage metadata

When the company grows, controlling metadata in a spreadsheet stops working. That's where metadata management and data catalog platforms come in. The best known in the market are Collibra and Alation, both very strong in governance and cataloging. Microsoft Purview is gaining ground in companies already living in the Microsoft and Azure ecosystem. Informatica brings metadata management integrated into its quality and integration suite. Ataccama and Atlan also show up more and more in this landscape.

One reminder I always repeat. No tool creates good metadata on its own. It catalogs, organizes, connects, and automates what the company already knows. If the company hasn't defined what a customer is, the software will simply record the mess in a prettier way. Technology speeds up governance, but it doesn't replace the human decision about what the data means. In the master data world this applies even more, because a material or supplier master without quality metadata is a master nobody truly trusts.

Metadata, data catalogs, and AI

The data catalog is, in essence, a big organizer of metadata. It works like an index of the company, a place where a person searches "revenue" and finds out where the data lives, what it means, who owns it, and whether it can be trusted. Without metadata, the catalog is empty. With metadata, it becomes the map that makes people stop reinventing reports that already exist.

And then there's artificial intelligence, which nobody ignores today. Enterprise AI depends on context to get things right. A model that doesn't know the difference between "gross amount" and "net amount," or that treats two records of the same supplier as different companies, will give a wrong answer with total confidence. Metadata is the context AI needs to understand the business. Without it, AI speeds up bad decisions, because it learns from a base it doesn't understand. Metadata governance is one of the quietest and most important foundations for AI to work inside the company.

How to learn to work with metadata

The study path goes through understanding data management broadly, not metadata in isolation. DAMA's body of knowledge, the DAMA-DMBOK, treats metadata as one of the core areas of data management, alongside quality, governance, and master data. The CDMP certification, tied to DAMA, is the most recognized for anyone who wants to build a serious career in the field. Beyond theory, real learning comes from getting your hands on a messy database, trying to document fields, putting together a simple glossary, and feeling firsthand why it's hard and why it's worth it.

Career opportunities

Metadata management is one of the most valued and least contested skills in data, and that creates opportunity. Everybody wants to be a data scientist, few want to organize the meaning of data, and that is exactly the work that holds everything else up. Professionals who know how to document, catalog, and govern metadata are rare and increasingly in demand, especially with the rise of AI and compliance requirements such as LGPD (Brazil's General Data Protection Law). Data steward, governance analyst, data catalog specialist, and data architect are roles that pass directly through this knowledge, and they are growing functions in the Brazilian market.

Main questions about metadata

Is metadata the same thing as data? No. Data is the information itself, such as a value or a name. Metadata describes that data, saying what it means, where it came from, and how it should be used.

Does metadata have anything to do with LGPD? Yes. To protect personal data, the company needs to know where it is, what it means, and who has access. That is metadata. A good inventory of personal data is, in practice, a metadata job.

Do I need an expensive tool to get started? Not to get started. You can begin with a glossary and a data dictionary in a spreadsheet, defining the most critical terms of the business. A tool makes sense when volume and the number of sources make manual control unworkable.

What's the difference between metadata and a data catalog? Metadata is the information that describes the data. A data catalog is the platform that gathers, organizes, and makes that metadata searchable for the whole company.

Who is responsible for metadata? Defining the meaning belongs to the business area, represented by the data owner. Day-to-day upkeep belongs to the data steward. Technology and method belong to the data area. It's shared work.

MDM Academy

MDM Academy is 4MDG's school of governance and master data, created to train professionals who solve real data problems inside companies. The idea is to translate concepts such as metadata, quality, master records, master data, and governance into practice applicable to the corporate world, in the language of someone who has lived through real projects. Those who study there learn to organize master records, support decisions with reliable data, and build a career in a field that only keeps growing. If you want a practical understanding of how to make governance work, it's worth getting to know MDM Academy and joining the Data Club, where the community exchanges experience and grows together.

Conclusion

Metadata doesn't show up in the director's report, doesn't have a pretty dashboard, and rarely gets credit when it goes well. But it decides whether the company can find, understand, and trust its own data. The paradox is that the most invisible information is what keeps everything else standing. Investing in metadata is investing in the trust of the entire base, and that is one of the data decisions that pays off the most over time. The question that remains is simple: at your company, when someone finds a number, can they tell where it came from and whether they can trust it?