Metadata

Types of metadata

Type	What it records	Examples for a product photo
Descriptive	What the asset is and what it shows	Title, caption, product name, product line, campaign, keywords
Technical	Properties of the file itself	Format, dimensions, resolution, colour space, duration, page count
Administrative	Who made it and who may use it	Creator, copyright, licence, model release, expiry date, approval status
Structural	How assets relate to each other	A product shot linked to its colour variants, a sales deck linked to its PDF export

Technical metadata is read from the file and needs no manual work. Descriptive and administrative metadata take the effort, and they carry most of the value: a sales manager looking for "the blue chair, outdoors, approved for Germany" depends on three fields that a person or a piece of software had to fill in.

Embedded metadata and standards

Some metadata travels inside the file. Cameras, editing software and photographers write it, and a DAM reads it when the file is uploaded.

• Exif (CIPA DC-008) records camera data: capture date and time, camera model, exposure settings and GPS position.
• The IPTC Photo Metadata Standard defines editorial and rights fields: headline, description, keywords, creator, copyright notice, credit line, usage terms, city and country.
• XMP (ISO 16684-1) is the format most current software uses to store IPTC and other fields inside images, PDFs and design files.
• Dublin Core (ISO 15836) is a set of 15 general elements (title, creator, subject, date, rights and others) used by libraries and archives, and a common reference for DAM schemas.

Ask photographers and agencies to fill in the IPTC creator, copyright notice and usage terms before delivery. Those values then arrive with the file instead of in an email that nobody can find a year later.

Designing a metadata schema

A metadata schema is the set of fields a library uses, with the type and allowed values of each. The usual mistakes are too many fields and too much free text. A schema that works for a marketing team usually has these properties:

• A short required set. Asset type, product or product line, campaign and usage rights cover most searches. People complete a short required set reliably and leave a long one half empty.
• Picked values instead of typed values for anything finite: regions, product lines, asset types, licence types. Free text drifts, so "packshot", "pack shot" and "product photo" become three values that a filter treats as different. A controlled vocabulary prevents this.
• Different fields for different kinds of asset. A product photo needs a product and a photographer, a sales deck needs an owner and a review date, a video needs a language.
• Fields that people search by. A field nobody filters on is rarely kept up to date. Check the search logs before adding one.
• Automatic values wherever possible, read from the file, the filename, the upload context or AI analysis, so that people only enter what software cannot know.

The DAM metadata best practices guide covers schema design in more detail.

Metadata in Freedam

• Exif and IPTC fields embedded in images are read on upload: description, keywords, creator, copyright, capture date and GPS position fill the asset's fields without typing. See ingestion.
• Every image is analysed by AI on upload, which adds tags, a caption and the text printed in the image. These are kept apart from the fields people fill in. See auto-tagging.
• Administrators define the fields for each asset class (text, number, date, yes/no or vocabulary), with required fields, default values, help text and values in several languages. Assets with a required field missing are flagged in the asset library.
• Vocabulary fields can depend on each other: once a product line is picked, the product field only offers the products in that line. See taxonomy.
• Filename parsing rules turn naming conventions such as SKU_colour_angle.jpg into field values on upload.
• Metadata can be edited on many assets at once, in a spreadsheet-style grid or with a bulk edit, and automation rules set fields when assets are uploaded or changed.
• Vocabulary fields can be shown as quick filters next to the gallery search bar, and a field such as a part number can accept a pasted list of values to find many assets in one search. See search.
• Access rules select assets by their metadata, so a field such as region or licence type can decide who may view or download an asset.

Frequently asked questions

What are examples of metadata?

For a photo: the caption, keywords, photographer, copyright notice, capture date, GPS position, dimensions and licence expiry date. For a document: the title, author, page count, language and review date. For a video: the duration, language, subtitles and the people who appear in it.

What is the difference between Exif and IPTC metadata?

Exif is written by the camera and describes how the photo was taken: date, camera, exposure, location. IPTC is written by people and describes what the photo shows and how it may be used: caption, keywords, creator, copyright and usage terms. Most images carry both.

Who should add metadata to digital assets?

Software should add everything it can: technical data, embedded fields, AI tags and values from filenames or a product catalogue. People add what only they know, such as the campaign, the approval status and the usage rights, ideally once at upload rather than later.