A DAM is where an archive's pictures are found, described and sent out. Whether it also keeps them safe is a separate question, and the answer is usually no.
The most important feature of a digital asset management system is the one nobody demonstrates: how you get everything out of it again. The Digital Preservation Coalition's Handbook observes that storage systems and media often last only three to five years. Software platforms last longer than disks, but not as long as photographs, and an archive that expects to exist in thirty years should expect to change systems several times. Every other requirement, from search to sharing, sits on top of that one.
This guide sets out what a photographic archive should ask of a DAM. It names no products, because products change faster than requirements, and because the right answer for a picture desk handling live news is the wrong one for an estate with fifty thousand transparencies.
What does a DAM do for a photo archive?
In working terms, five things.
- Ingest. It takes in files, reads the metadata already inside them and makes previews.
- Describe. It holds captions, keywords, dates, names, credits and rights notes against each image or group.
- Find. It searches all of that, and lets users save selections as lightboxes or collections.
- Deliver. It produces copies at the right size and format, with or without watermarks, and shares them with people outside the organisation.
- Record. It keeps a history: who downloaded what, which licence covered which use.
A DAM is the archive's front desk. It is what a picture library shows its clients, and for many archives it doubles as the public website.
Is a DAM the same as a preservation system?
It is not, and confusing the two is the most expensive mistake in this field. A DAM is designed around access. Preservation is designed around the assumption that any one system will fail.
The Digital Preservation Coalition's guidance on storage asks for multiple independent copies in different geographic locations, on different storage technologies, with fixity information such as checksums used to record and regularly monitor the integrity of each copy. A DAM hosted by one supplier in one cloud is, on that test, a single copy. It may be a well-protected one. It is still a system that can lose your account, change its terms or cease trading.
The safe arrangement is a division of labour. Preservation masters live under a 3-2-1 backup routine with checksums that the archive controls. The DAM holds access copies, or points at the masters, and can be rebuilt from them if it has to be. If a system offers fixity checking and replication, treat that as a welcome addition and not as a substitute.
Which requirements matter most?
| Requirement | Why it matters | How to test it |
|---|---|---|
| Complete export | You will leave one day | Export a trial set: originals untouched, all metadata in CSV or XML |
| Originals preserved bit for bit | A master that has been re-saved is no longer the master | Compare checksums before upload and after download |
| Embedded metadata, read and written | Descriptions must travel with files that leave the system | Edit a caption, download the file, inspect it elsewhere |
| Format support | Archives hold raw files and large TIFFs, not only JPEGs | Upload your largest scan and a raw file from each camera model |
| Hierarchy | Photographs belong to folders, jobs and collections | Model one real box: collection, folder, frame |
| Your identifiers | Catalogue references must survive | Search by archive reference; check it appears on export |
| Controlled vocabularies | Consistent terms make search reliable | Load a term list; try to enter a term that is not on it |
| Rights and usage fields | Licensing needs restrictions and history against each image | Record a restriction and see whether it blocks or warns on download |
| Permissions | Researchers, clients and staff need different views | Create an external user and check what they can see |
| Duplicate handling | The same picture arrives many times | Upload a file twice, then a resized copy |
| Open interfaces | Other tools will need to talk to it | Ask for the documentation of the programming interface |
Three of these deserve more than a row.
Embedded metadata. The Library of Congress recommends that photographs carry their descriptive information in the file where the format allows, using a common embedded schema such as IPTC. A system that keeps captions only in its own database produces files that are anonymous the moment they are downloaded. The IPTC fields every archive should fill are the minimum a DAM must round-trip without loss. You can check any downloaded file with the EXIF and metadata viewer, and write the core rights fields into JPEGs with the tool to add copyright metadata if a system leaves them out.
Hierarchy. Archival description runs from the whole collection down through series and files to items, as The National Archives' cataloguing guidance explains. Many DAMs are flat: every asset is a peer of every other, grouped only by tags. For a photographer's archive that loses the single most useful fact about a frame, which is what folder and job it came from. If the system cannot model hierarchy, the archive reference in each record has to carry it.
Duplicates and variants. Archives accumulate the same image as raw file, TIFF, retouched version and several JPEGs. A good system links these as versions of one photograph. A poor one presents five results for one picture. The mechanics are covered in the guide to finding duplicates with hashes.
What should you ask about automatic tagging?
Most systems now offer automatic keywording and visual search. They can be useful for a first pass through uncaptioned material, and they are a poor source of names, dates and places, which are what researchers search for. The strengths and limits are discussed in the article on AI image tagging for archives. Two questions belong in any procurement:
- Are machine-generated tags stored separately from human cataloguing, and labelled as such?
- Are your images used to train the supplier's models, and can that be switched off in the contract?
The second is a rights question as much as a technical one. An archive that licenses its images on negotiated terms should not find that its own software supplier has acquired a use of them by default.
How should you run a trial?
- Write requirements before watching demonstrations. Rank them: essential, useful, irrelevant.
- Prepare a test set of a few hundred files that reflects the real archive: raw files, large TIFFs, JPEGs with and without metadata, a few duplicates, one complete folder with its catalogue records.
- Record checksums for the test set before it goes anywhere.
- Load it yourself. How long it takes, and what the system does to file names and metadata on the way in, is the truest part of any trial.
- Search as a stranger would. Ask someone outside the project to find ten specified pictures.
- Export everything and compare. Files identical, metadata complete, identifiers intact.
- Price the whole life. Storage growth, user licences, support, and the cost of leaving.
Can you start without a DAM?
Yes, and most archives should. The groundwork is the same in either case: a folder-level catalogue, unique file names, metadata embedded in the files and verified backups. An archive that has those can adopt a system in weeks. One that lacks them will pay a supplier to discover the fact.
The Robert Harper Photograph Library is at that deliberate, pre-system stage in public. Its inventory of 503 folders is published as a searchable page with a CSV download at /archive-inventory/, and 330 described collections are browsable by category, from fashion to marine. The inventory counts 95,487 unique photographs, or 177,143 frames including duplicates and variants, held digitally as Nikon NEF raw files, TIFFs and JPEGs, with the film, mainly 6 x 7 colour transparency, largely unscanned. The 503 inventory lines add up exactly to those totals, though they have not been checked frame by frame against the physical material. There is one rights holder, 825 Ventures Ltd, which simplifies the rights fields any future system will need.
A plain list of that kind answers the first question any researcher asks, which is whether the archive has anything on their subject. Picture editors, agencies and researchers who want to query it, or to discuss how the collection could be delivered into their own systems, will find the details on the page for picture agencies.
Frequently asked questions
What is a DAM system?
A digital asset management system is software that stores media files with their descriptions and lets people search, group, convert and share them. For a photo archive it is the working catalogue: the place where a researcher types a subject and gets pictures back, and where captions, keywords, credits and usage records are kept against each image.
Does a small photo archive need a DAM?
Not at first. A consistent folder structure, unique file names, a spreadsheet inventory and embedded metadata will carry an archive a long way, and they are the preparation any DAM needs in any case. A system starts to earn its cost when several people need to search, when images go out to clients regularly, or when rights and usage have to be tracked.
Is a DAM the same as a digital preservation system?
No. A DAM is built for finding and delivering files. Preservation means keeping independent copies in different places, checking them against recorded checksums and migrating them as storage ages. Some systems offer parts of this, but an archive should assume its DAM is the access copy and manage preservation masters under a separate, documented routine.
What metadata standards should a photo DAM support?
It should read and write the IPTC Photo Metadata Standard, stored in the file as XMP, and preserve the EXIF data cameras record. Those are the fields picture desks, search engines and other libraries read. It should also let you define your own fields for archive references, physical locations and rights notes, and map them on export.
How do I avoid being locked in to a DAM vendor?
Test the exit before you sign. Ask for a full export of a trial collection: original files, unchanged, plus all metadata in an open format such as CSV or XML, with your identifiers intact. Keep descriptive metadata embedded in the files as well as in the database, and keep your own inventory outside the system.
Sources
- Digital Preservation Coalition Handbook: Storage
- Digital Preservation Coalition Handbook: Fixity and checksums
- Library of Congress: Recommended Formats Statement, still image works
- The National Archives: Cataloguing archive collections
Published by Archive.London. Corrections are welcome under our editorial policy.

