A digital asset management platform earns its place at the moment somebody finds a file they would otherwise have recreated. Everything else — storage, structure, governance — is in service of that. Finding works in two distinct modes, and knowing which one you are in makes the difference between locating something in seconds and giving up.
Search starts from the bar at the top of the platform and returns suggestions as you type. Each suggestion shows the asset name, a thumbnail, the folder it lives in, and — usefully — the metadata field the match was found in, so you can tell whether the hit came from a caption, a product code or a photographer's name.
The search covers more ground than a filename lookup:
That last one is worth pausing on. A phrase appearing on page nine of a PDF makes that document findable even though nobody recorded the phrase as metadata, which is often how a report or contract gets located when no one remembers what it was called.
Pressing Enter opens the full results page, where you can search assets alone or across all record types, and sort by name, extension, type, creation date or modification date.
Search results are ranked by how closely each asset matches, and precision sets the threshold for inclusion. It defaults to 40 per cent, deliberately loose enough to tolerate approximate recollection.
The control matters because the two ways a search fails have opposite fixes. Too many results means the threshold is too low — raise precision and the list shortens to close matches. Nothing useful means the opposite: lower it and looser matches come into range, which is what you want when you half-remember a term or are not sure what the asset was called. Reach for precision before filters when results feel wrong, since it is one adjustment rather than a set of decisions.
Filters answer a different question. Rather than describing a specific asset, you are describing a category — everything shot in Berlin, everything approved for external use, every video added since March — and reducing the library to it. Filters are available in the library, the metadata view and search results, and they can be applied on top of a search to narrow what it returned.
Quick filters cover the common criteria: metadata profile, metadata values, tags, file extension, copyright, and creation or modification dates. When a field has more values than can sensibly be listed, a search box inside that field looks any of them up directly, so you are never limited to the shortlist shown by default.
Advanced filters add conditional logic. Conditions can be combined with AND, OR, or custom expressions such as 1 AND (2 OR 3), and can test folder location, subfolder contents, preview availability and copyright alongside metadata. This is where genuinely specific questions get asked — approved product photography from two particular campaigns, excluding anything expiring this quarter.
A filter worth constructing is usually worth keeping, and there are two ways to keep one.
A saved filter is a reusable configuration applied in a click. It lives in the filter panel and under Library > Filters, and administrators can share saved filters with other users, which is how a team ends up with a common set of views instead of everyone rebuilding the same criteria. Saved filters also reach beyond browsing: a saved filter can drive a showcase, controlling which assets appear on a public gallery.
A smart folder takes the same criteria and presents them as a folder in the library structure, populated automatically by anything matching. Assets appear and disappear as their metadata changes, without anyone maintaining the contents. Smart folders can be nested like ordinary folders, with two limitations worth knowing: you cannot upload directly into one, and they cannot be added to a portal.
The choice comes down to whether you want a view you apply or a place you visit. Use a saved filter to repeat a search; use a smart folder when people should find the set by browsing, without knowing a filter exists.
Folders, search and filters are not competing organisational schemes, and treating them as one is the usual cause of an over-engineered folder tree. Folders are where an asset lives, and they carry permissions. Search is how you find a specific thing you can describe. Filters are how you isolate a set defined by shared properties.
Once search and filters do their job, the folder structure stops needing to anticipate every way someone might look for something — which is fortunate, because no folder tree ever manages it. A structure organised around ownership and access, with finding delegated to search and filters, is easier to maintain and easier to explain.
Search matches what has been recorded. An asset with a camera-generated filename, no description and no tags is findable only by someone who already knows where it is, and no amount of tuning changes that.
This is the practical argument for metadata discipline, and it is more persuasive than any appeal to good practice. Every field filled in is another route to the asset: a campaign name, a product code, a location, a usage right. When people complain that they cannot find anything, the cause is almost never the search — it is that the library was filled faster than it was described.
If results are consistently disappointing, look at what a typical asset actually carries. Fields that are always empty are fields nobody can search on, and either the field should go or the workflow that skips it should change.