Metadata is data that describes other data. It tells you about a piece of information without being the information itself.
Think of a library card. The card lists a book's title, author, publication year, and shelf location. None of that is the book's actual content—it's metadata that helps you find and understand the book.
According to NISO's metadata primer, metadata is "structured information that describes, explains, locates, or otherwise makes it easier to retrieve, use, or manage an information resource." This definition highlights three key functions: description, location, and management.
Metadata itself never replaces the original content. A song's metadata (artist, album, duration) is not the music. A photo's metadata (camera model, date taken) is not the image. The metadata describes; the data is.
Metadata vs. Data: What's the Difference?
Data is the content you create or consume. Metadata is the description attached to it.
Consider a photo on your phone. The image you see—the pixels, colors, and scene—is the data. The camera model, shutter speed, GPS coordinates, file size, and timestamp stored alongside it? That's metadata.
| Aspect | Data | Metadata |
|---|---|---|
| Definition | The content itself | Information about the content |
| Photo example | The image pixels | Camera, date, resolution, location |
| Document example | The text you wrote | Author, file type, word count, version |
| Size | Usually larger | Usually smaller |
| Format | Often unstructured | Typically structured |
Metadata tends to be smaller and more consistently formatted than the data it describes. This makes it easier for systems to index, sort, and search.
Understanding the difference between structured vs. unstructured data clarifies why metadata is so useful: it adds structure to otherwise messy content.
The Main Types of Metadata
Most metadata falls into three categories: descriptive, structural, and administrative. Each serves a different purpose.
| Type | Purpose | Examples |
|---|---|---|
| Descriptive | Helps you find and identify content | Title, author, keywords, abstract |
| Structural | Shows how content is organized | Chapters, page order, file relationships |
| Administrative | Supports management and governance | Creation date, file format, permissions, provenance |
Some classification systems add preservation metadata (for long-term archiving) and rights metadata (licensing, usage terms). These often fall under the administrative umbrella.
Descriptive Metadata
Descriptive metadata helps you find and identify a resource. It answers questions like: What is this? Who made it? What is it about?
Key point: Descriptive metadata powers search and discovery.
Common examples include:
- Title: The name of a document, video, or page
- Author/Creator: Who produced the content
- Keywords/Tags: Terms that describe the subject
- Abstract/Summary: A brief description of the content
A book's title and author name are descriptive metadata. On a web page, the <title> tag and meta description serve the same function.
Structural Metadata
Structural metadata describes how parts of a resource relate to each other. It answers: How is this organized?
Key point: Structural metadata maps the internal organization of content.
Examples include:
- Chapters and sections in a book or report
- Page order in a scanned document
- Track sequence in an album
- Relationships between files in a dataset
If you've ever used a table of contents or skipped to a specific chapter in an e-book, you relied on structural metadata.
Administrative Metadata
Administrative metadata supports the management, preservation, and governance of resources. It answers: When was this created? Who can access it? Where did it come from?
Key point: Administrative metadata tracks provenance, rights, and technical details.
Examples include:
- Creation date and modification history
- File type and format (PDF, JPEG, MP4)
- Permissions and access controls
- Provenance (origin and chain of custody)
U.S. federal law (44 U.S.C. § 3502(19)) defines metadata as "structural or descriptive information about data such as content, format, source, rights, accuracy, provenance, frequency, periodicity, granularity, publisher or responsible party, contact information, method of collection, and other descriptions." This legal definition shows how broadly administrative metadata can extend.
Metadata Examples in Everyday Life
Metadata surrounds you. Here are examples you interact with regularly:
- Photo (EXIF data): Camera model, shutter speed, aperture, ISO, GPS coordinates, date and time taken. Your phone stores all of this every time you snap a picture.
- Document (Word, PDF): Author name, creation date, last modified date, word count, version number. These fields appear in file properties.
- Email: Sender, recipient, timestamp, subject line, routing headers. The message body is data; everything else is metadata.
- Music file (ID3 tags): Artist, album, track title, genre, year, album art. Music players use these tags to organize your library.
- Web page: Title tag, meta description, Open Graph tags, canonical URL, author. Browsers, search engines, and social platforms read this metadata.
Each example follows the same pattern: the metadata describes the content without being the content.
Key point: You already use metadata constantly—you just might not call it that.
Why Metadata Matters
Metadata makes data findable, understandable, trustworthy, and reusable. Without it, information becomes a pile of unlabeled files.
Key point: Good metadata turns raw content into organized, searchable, and actionable information.
Organizations use metadata for:
- Search and discovery: Finding the right file, document, or record
- Organization: Grouping content by type, date, author, or project
- Governance and compliance: Tracking who created what, when, and who can access it
- Interoperability: Ensuring systems can exchange and understand data
- Trust: Verifying source, provenance, and accuracy
The business value is substantial. According to estimates by Grand View Research, the global metadata management tools market is projected to reach USD 36.44 billion by 2030, expanding at a CAGR of 20.9% from 2025 to 2030.
For AI systems, metadata is especially critical. It tells a model where information came from, when it was published, and whether it can be trusted. Raw text alone cannot provide this context. An AI answering questions about current events needs to know whether a source was published yesterday or five years ago—that's metadata doing its job.
Metadata on the Web: Meta Tags and Structured Data
Web pages embed metadata directly in their HTML. This metadata tells browsers, search engines, and other tools what the page is about.
Key point: Web metadata lives in the page's <head> section and in structured data markup.
Common types include:
- HTML
<meta>tags: Title, description, viewport, character set, robots directives - Open Graph tags: Control how the page appears when shared on social platforms (image, title, description)
- JSON-LD / Schema.org: Machine-readable structured data describing products, articles, events, organizations, and more
An HTML parser reads this markup and converts it into a structured format that programs can process.
Structured metadata adoption has grown rapidly. The HTTP Archive Web Almanac 2024 reports JSON-LD adoption rose from 34% of pages in 2022 to 41% in 2024, while Open Graph metadata appears on 64% of pages. Search engines, social platforms, and AI systems now expect pages to include machine-readable metadata.
How Metadata Powers Web Data Extraction and AI
When you scrape a web page, you capture more than just the visible text. You also extract metadata: the source URL, page title, author, publish date, HTTP status code, and provenance.
A web data extraction API returns both the content and its metadata as structured fields. Tools that convert HTML into structured JSON include metadata alongside the extracted content.
Key point: Metadata gives AI systems the context they need to trust and attribute information.
For AI applications—especially retrieval-augmented generation (RAG) pipelines—metadata serves several functions:
- Source attribution: Which URL did this content come from?
- Freshness: When was it published or last updated?
- Authority: Who is the author or publisher?
- Filtering: Can the system exclude outdated or low-quality sources?
Schema-based extraction can extract structured data from HTML, pulling defined metadata fields even when page layouts change. This approach is more resilient than relying on fixed CSS selectors.
The scale of embedded metadata on the web is significant. According to WebDataCommons 2024 analysis, structured data was found within 51.25% of the roughly 2.4 billion web pages examined in 2024, up from just 5.7% in 2010.
Proper web data parsing captures this metadata and organizes it into fields your application can use directly. Without metadata, web-extracted content is just text—useful, but missing the context that makes it trustworthy and actionable.
How to View or Find Metadata
Finding metadata is straightforward once you know where to look.
Files (photos, documents, videos):
- Windows: Right-click the file → Properties → Details tab
- macOS: Right-click → Get Info, or open in Preview and choose Tools → Show Inspector
- EXIF viewers: Specialized tools display detailed photo metadata (camera settings, GPS, etc.)
Web pages:
- View Source: Right-click → View Page Source, then look for
<meta>tags in the<head>section - Developer tools: Open your browser's DevTools (F12), go to the Elements tab, and inspect the
<head> - Structured data testing tools: Google's Rich Results Test shows JSON-LD and schema markup
Organizations:
- Data catalogs: Enterprise systems like data catalogs aggregate metadata across databases, files, and APIs
- Database schemas: Column names, data types, and relationships are structural metadata
Frequently Asked Questions
What Is a Simple Example of Metadata?
A photo's date taken, camera model, and GPS coordinates are metadata. They describe the image without being the image itself.
What Are the Three Main Types of Metadata?
Descriptive metadata helps you find content (title, author). Structural metadata shows how content is organized (chapters, sequence), while administrative metadata supports management—creation date, permissions, and provenance.
What Is the Difference Between Data and Metadata?
Data is the content itself—the text, image, or recording. Metadata describes that content: its title, source, format, and other attributes.
What Is Metadata Used For?
Metadata helps you find, organize, understand, and trust data. It powers search engines, enables data governance, and gives AI systems the context they need to attribute sources.
What Is Metadata in a Web Page?
Web page metadata includes HTML <meta> tags (title, description), Open Graph tags (for social sharing), and structured data (JSON-LD, schema.org) that describe the page to browsers, search engines, and crawlers.
