What is tagging?
Last updated: August 2026
Tagging means assigning descriptive keywords (tags) to digital content such as images, videos or documents. Tags describe what a file shows or is about, making it searchable and filterable by meaning rather than by file name. Tagging — also called keywording — is the foundation of findability in media libraries, DAM systems and document management systems.
What does tagging mean in practice?
To tag a file is to attach the terms under which people will later look for it. Instead of identifying an image only by its name (“IMG_4711.jpg”), tags describe its content: “power drill”, “product shot”, “cut-out”, “autumn campaign”. Anyone searching for one of these terms finds the file — regardless of what it is called or where it is stored.
Tagging is an investment: the effort occurs once at capture, the payoff at every later search. The reverse also holds — an untagged asset in a large library is effectively invisible.
What is the difference between tags, keywords and metadata?
Tags and keywords largely mean the same thing: terms describing a piece of content, assigned freely or from a controlled list. “Keywording” is the traditional term in photography and stock media; in marketing, “keyword” additionally refers to search terms for SEO and advertising — a related but separate topic.
Metadata is the umbrella term: all structured information about a file. Tags are one part of it, alongside technical data (resolution, colour space, EXIF camera data), descriptive fields (title, caption) and administrative data (creator, usage rights, expiry dates). Standards such as IPTC and XMP embed this information in the file itself, so it travels between systems and can be read by any compliant application — a DAM system imports embedded metadata and makes it searchable.
How do you tag images and documents correctly?
Good tagging follows rules, not gut feeling. The essentials:
- Define a vocabulary first: An agreed keyword list (or taxonomy) prevents sprawl like “car/auto/vehicle” for one concept.
- Standardise forms: Singular or plural, capitalisation, language — decide once, apply consistently.
- Think like the searcher: Tag what colleagues will search for, not what comes to mind while uploading.
- Cover several dimensions: What is shown (subject)? What is it about (topic)? What is it for (use, campaign, channel)? Which rights and deadlines apply?
- As many as needed, as few as possible: A handful of precise tags beats a long arbitrary list — every wrong tag produces wrong search results.
Controlled vocabulary vs. free tagging — which is better?
For teams and growing libraries, a controlled vocabulary (taxonomy) is almost always the better foundation; free tags complement it where flexibility is needed.
| Criterion | Controlled vocabulary / taxonomy | Free tagging (folksonomy) |
|---|---|---|
| Consistency | High — everyone uses the same terms | Low — synonyms and spelling variants accumulate |
| Maintenance | Vocabulary needs curation and extension | No setup, but growing sprawl |
| Search quality | Reliable, complete results | Depends on the discipline of each contributor |
| Flexibility | New terms require a process | Instantly extensible, good for trends and projects |
| Best for | Teams, large libraries, DAM/DMS/PIM | Individual users, small collections, transitions |
| Best practice | Core taxonomy as the mandatory base … | … plus a limited set of free tags on top |
Related but distinct: classification assigns content to exactly one position in a fixed hierarchy (such as product classification standards), whereas an asset can carry any number of tags. See classification of product data for the product data side.
How does AI-based auto-tagging work?
AI-based tagging analyses the content of a file and suggests appropriate tags automatically. Image recognition models identify subjects, objects, colours and scenes; language models extract topics and key terms from documents; speech-to-text makes audio and video content searchable.
The proven setup is two-staged: AI handles the bulk work and proposes tags, while a person spot-checks and adds what only context knowledge can supply — campaign membership, rights status, internal product names. This cuts capture effort substantially without leaving quality to chance. The controlled vocabulary remains essential: AI suggestions should be mapped to your own taxonomy, otherwise you automate the sprawl.
Why is tagging important?
Tagging decides whether digital content creates value or merely occupies storage. Without it, assets are reproduced because nobody can find them; with it, libraries become reusable, rights become checkable and processes measurably faster: search times drop because results match meaning instead of file names, duplicate production disappears, usage rights and expiry dates live on the asset itself, and downstream systems — from web shop to print catalogue — can filter and publish content automatically.
What is tagging in a DAM system?
In a DAM system, tagging is a core capability: media files — images, videos, graphics, layouts — are tagged with a focus on subject, usage, brand and rights. Document management systems apply the same principle to business documents, focusing on document type, process and retention. In product data environments, a third layer matters: in a PIM system, tags and attributes connect products with their media, so that, for example, all cut-out images of a product line are automatically assigned to the right article and channel.
Where does tagging sit in the PIM/PXM/DAM landscape?
Tagging is a core function of digital asset management: it makes the media library findable and controllable. In combination with a PIM system, tags and attributes link media to product data — the prerequisite for PXM processes that deliver channel-specific product experiences. Keeping classification, attributes and tags cleanly separated lays the foundation for consistent product communication.
How does a DAM system like OMN support tagging?
In the DAM module of OMN (Online Media Net), assets are enriched centrally with tags and metadata: controlled vocabularies and mandatory fields safeguard consistency, embedded metadata (e.g. IPTC/XMP) is imported automatically, and the link to product data in the PIM ensures media lands on the right article and channel. Workflows govern who tags and who approves. OMN’s DAM also covers the bulk work: with the AI Tagging add-on, AI services analyse your images right after upload, recognise subjects, objects and scenes and suggest matching tags — including OCR text recognition and automatic translation of tags into your preferred language. You review and approve the suggestions in your usual workflow. apollon contributes more than 25 years of experience in managing media and product data. For a market overview, see the DAM software comparison — or request a demo to see OMN’s tagging workflows in action.
FAQ — Frequently asked questions about tagging
How many tags per asset are reasonable?
As a rule of thumb, about 5 to 15 precise tags per asset: enough to cover subject, topic and usage, few enough to avoid false hits. The number matters less than the rule behind it — every tag must answer a realistic search query.
Who should do the tagging?
Ideally the role that knows the context: media producers for subject and technical aspects, product management for product relations, marketing for campaigns. A short tagging briefing plus a controlled vocabulary keeps quality stable across roles.
Can tags be corrected later?
Yes. DAM and DMS systems allow tags to be added, replaced or bulk-renamed at any time. When migrating vocabularies, map old terms to new ones so existing searches and automations keep working.
What is the cost of missing tags?
It shows up as search time, duplicate production and rights risks: assets nobody finds are recreated or used unchecked. The effort of clean tagging is consistently smaller than the follow-up costs of an unfindable library.