Data classification

Data classification is the grouping of files by what they are, from metadata signals rather than content, such as name, path, owner, and structural detail. It groups files so a rule can be written once and applied to the group.

Key takeaways

Data classification groups files using metadata signals such as name, path, age, and owner.
It works without reading file contents, which keeps it fast and non-invasive at scale.
Classification produces the labels, and governance decides what happens to each group.

Why does data classification matter?

Because rules can then be written once per group instead of once per file. When files are sorted into groups such as crown jewels, rebuildable files, sensitive records, and waste, each group gets one decision about protection, storage tier, retention, or deletion. Without classification, every file looks the same, so either everything is kept on expensive storage or nothing is protected well enough.

How is data classification done without reading file contents?

Classification from metadata uses signals the file system and the index already hold, such as name, path, owner, and structural detail supplied by format-aware plugins. It groups files by what they are and where they sit. It does not detect what a document says, and it is faster and less invasive because of that.

What is the difference between classification and governance?

Classification sorts data into groups. Governance decides what happens to each group and proves it happened. Classification produces the labels, and governance acts on them and records the result. Both steps are needed, and each is incomplete alone.

How do you classify data using metadata?

By turning what people already know about their files into rules. The usual steps:

Choose the groups. Start from decisions the business needs to make, such as what to protect, what to move, and what to delete, rather than from file formats.
Find the signals for each group. Look for attributes such as file extension, folder pattern, owner group, and size that reliably mark a group. For example, graded video masters are often .mxf, .mov, or .exr files in folders named /graded/ or /final/.
Write the rules. Turn each set of signals into a rule that tags matching files with their group.
Check the results. Review a sample of each group with the people who own the data, and adjust the rules where files land in the wrong place.
Run the rules on every scan. New files are then classified as soon as they appear.

What are common data classification categories?

Most organizations classify the same files in more than one way, because each view answers a different question:

By value: files that cannot be recreated, such as original recordings or signed contracts, apart from files that can be rebuilt, such as proxies, renders, and caches.
By sensitivity: levels such as public, internal, confidential, and restricted, which decide who may see the data and where it may be stored.
By lifecycle stage: active, inactive, under retention, or ready for deletion.
By ownership: the project, department, or cost center the data belongs to.

What happens to files that no rule matches?

They stay in an unclassified group, and that group is useful in its own right. Its size shows how much of the estate is still unexplained, and the largest folders and owners inside it show where to write the next rule. A shrinking unclassified group is a simple way to measure progress, and a growing one often points to a new project or system that nobody has described yet.

How does Diskover help?

Diskover classifies files from their metadata as it indexes them, and enrichment plugins add structural detail for many formats, so each group can be searched, reported on, and acted on.

Explore data intelligence

Book a demo