The fundamental paradox
Ask any marketing team what they expect from a DAM. The answer is always the same: “find our files in three clicks.”
Now ask them who will fill in the metadata: the title, keywords, rights, expiration date. Silence.
This is the central paradox of every DAM project. Everyone wants the destination. Nobody wants to pave the road.
Files, not assets
Theresa Regli, who has guided DAM strategy for over 20% of the Fortune 500, sums up the problem in one formula: “Content + metadata = an asset.”
Without metadata, you don’t have digital assets. You have files. Thousands of files in a tool that’s slightly more expensive than a shared Drive.
And the numbers confirm it. According to a Canto/Civey study, 33% of marketing professionals spend three weeks a year searching for images, videos and files on disorganized servers. McKinsey estimates that employees spend an average of 1.8 hours per day searching for information. This is not a tool problem. It’s a metadata problem.
Why nobody tags
The instinct is to blame laziness or lack of discipline. That’s more comfortable than admitting the truth: the system is poorly designed.
Three causes come up again and again.
The taxonomy is too complex. When uploading a photo requires filling in fifteen mandatory fields, people bypass the system. They drop the file on the nearest shared Drive. The DAM empties out while the chaos moves elsewhere.
Every team speaks its own language. Marketing says “campaign visual”, communications says “brand asset”, legal says “rights-managed document”. As metadata expert Yonah Levenson notes, using “sofa” instead of “couch” seems trivial. Until someone searches for “couch” and finds nothing.
The effort is invisible. The person who tags a file properly never gets thanked. The person who can’t find a file blames the tool. Metadata work is thankless by nature: when it’s done well, nobody notices.
The silent debt
In software development, we talk about “technical debt”, shortcuts that accumulate and eventually paralyze a system. Metadata has its equivalent.
Every file uploaded without a tag, every field left blank, every ignored convention adds a line to this debt. Individually, it’s negligible. Collectively, after six months, it’s an unusable search system, invisible duplicates, and rights that are impossible to verify.
67% of teams abandon their DAM within 18 months. The diagnosis is almost always the same: the tool works, but nobody can find anything. The fault doesn’t lie with the search engine. It’s that there’s nothing to search for, in any structured sense.
What AI won’t solve (on its own)
AI-powered auto-tagging is often presented as the silver bullet. And it does provide real help: image recognition, technical metadata extraction, keyword suggestions.
But AI recognizes what’s in a file, not what it means to your organization. It can identify a photo of a building. It doesn’t know it’s the headquarters, that the image expires in December, and that it should only be used internally.
Business context remains human. AI accelerates metadata work. It doesn’t eliminate it.
The real issue: making tagging viable
The problem isn’t that people refuse to tag. It’s that they’re asked to do it under impossible conditions.
DAM projects that work don’t start with a perfect metadata schema. They start with a minimal viable one:
- few mandatory fields at import (category, rights, date). The rest comes later;
- controlled vocabularies rather than free-text fields: you pick from a list, you don’t invent;
- progressive enrichment: metadata gets completed over time, not at the moment of upload;
- lightweight governance: someone owns the schema, reviews it regularly, and listens to users.
It’s less ambitious than an exhaustive data model. It’s far more realistic.
The question to ask
Not “how do we force people to fill in metadata?”
But: what is the minimum structure needed so that assets are findable, usable, and traceable?
The answer is always simpler than you’d think. And it’s what determines whether your DAM will be a working tool or a file graveyard.