Schemas and Taxonomies and Ontologies, Oh My!
A visual guide to simplify complex terms
By Erica Hornung
In a previous blog post, we discussed ways a DAM (Digital Asset Management) system could benefit your company. You’ve taken the next step and purchased a DAM – congratulations! – but now it’s time to set it up, and suddenly you need to learn new terminology. If you’re wondering, “what’s a schema? What’s the difference between a schema and a taxonomy? What the heck is an ontology, anyway?” don’t panic! While the formal definitions of these terms can be quite complex, with enough overlap to be confusing, the concepts are already familiar to you and your organization – and are easily understood by illustrating practical examples. A picture is worth a thousand words, as they say – let’s look at schemas, taxonomies, and ontologies.
Schemas: What do you want to know about your assets?
Complicated definitions aside, a schema is really just a list of things you want to know about your assets. Think of your schema as a form to fill out, with places to enter information about your assets that are pertinent to your organization. Your schema reflects the needs of your business and how you use your assets. For example, a digital asset schema for an advertising agency might include data points such as the type and size of an image, the usage rights, the campaigns an image was used for, and the date it was created (Figure 1).
When properly set up, each data point is searchable in a DAM, so a user might find the image on the left by searching for Theodore Roosevelt or Project Number 2015Q1P12. The image on the right can be found by searching by Campaign National Dog Day, or the keywords dog or cowboy.
Different institutions may want to know different things about their assets; for example, the Library of Congress has a different list of data points to collect - a different schema for its collection of digital images. Even your desktop Finder window has a schema for displaying image information (Figure 2).
All the metadata in the above schemas relate to the same image, but each use case has different metadata needs, and therefore a different schema in place. When designing a schema to use, an organization must consider what information is critical to its unique needs.
Taxonomies: How are my assets organized?
Once an organization has designed the form to fill out – the schema - for ingesting images into a DAM, they will organize those images to make them easier to find. The hierarchy in which assets are organized is called a taxonomy.
One familiar way to organize assets hierarchically is to create file folder structures to keep related assets together (Figure 3).
In a taxonomy, the top-level, or parent, folder can have any number of child folders, which in turn may have their own child folders. Each child folder, however, can only have one parent folder - the Campaigns folder in Figure 3 can have multiple child folders, each for a different type of campaign, but the Outdoors folder can only have one parent, Campaigns.
Organizing assets by campaign makes sense for an advertising agency – they need to keep all assets used in a campaign together. Another organization, however, might want to organize their digital images by color model – black and white images in one folder, color images in another folder.
Just like schemas, taxonomies are designed with the unique needs of the organization in mind. An organization needs to carefully consider which data point is most important to organizing assets – is it by rights status? File type? Photographer? Whatever that data point is, a taxonomy can be built to logically file assets for easy retrieval.
Ontologies: When taxonomies aren’t enough
What if an organization wants to organize assets in two different hierarchies – by Campaign and by Color? Or what if an asset is used in two different campaigns? What happens when it’s decided to group assets by Project Number, as opposed to Campaign name? Enter the ontology – managing the relationships between different taxonomies.
To locate the image of Theodore Roosevelt the folder structure shown in Figure 4, a user would navigate to the Campaigns/Holiday/Talk like a Cowboy Day/B_W folder… or is it the Campaigns/Holiday/President’s Day folder? What if we wanted to further sort our assets by file size, or rights status?
Due to its hierarchical nature, a taxonomy can quickly become overly complex and difficult to navigate and manage. Likewise, taxonomies don’t handle situations where one asset is used in more than one campaign very well; keeping multiple copies of assets in multiple locations is not only a waste of space, but can quickly become confusing – “which copy of this image was I working on?”
An ontological approach to digital asset management solves this problem by defining the relationships between assets and taxonomies. An advertising agency that wants to organize images by campaign and by rights management status can do so by designing a schema to include those data points.
By including the data point Campaigns in the metadata schema attached to an asset, assets used in multiple campaigns are easily organized and searchable in the DAM as part of either campaign. No more duplicate files in multiple folders! Simply searching the DAM by the Campaign field with the values Outdoors or President’s Day will return the same image of TR.
What if we want to organize assets by rights management status, regardless of the campaign it’s used in? By including the Usage Rights data point in our schema, it’s easy to organize assets by rights type.
By now you can see the incredible flexibility offered by using an ontological approach to designing a metadata schema – it’s possible to organize assets by any number of hierarchical relationships. The possibilities are limited only by your metadata schema!
Figure 6 shows the relationships between two image assets and multiple taxonomies: Campaigns, Keywords, Usage Rights, Asset Type, Asset Format, and Permissions. Imagine trying to keep track of all this information by organizing assets in folders!
What’s the bottom line for my business?
By designing a metadata schema specifically for the needs of your company, your assets will be easier to find and organize in your DAM than ever before, saving time and eliminating wasteful duplication, leading to a higher ROI on your valuable digital assets. Relying on this schema to organize assets into various taxonomies as needed- rather than relying on overly complex file folder structures – will leverage the ontological relationships between assets and ultimately make your DAM integration project a success.
Salt Flats’ team of library scientists and data managers has extensive experience in designing customized metadata schemas to unleash the ontological possibilities in DAM systems. Whether it’s in the Media & Entertainment, Health Care, or CPG industry and beyond, our subject matter experts can work with your company to design a metadata strategy to ensure your company gets the most value possible from your DAM. Ready to know more? Reach out to the Salt Flats team to start a conversation today.
Image credits:
Johnston, Frances Benjamin, photographer. Montauk Point, Rough Riders, Col. Roosevelt. Photograph. Retrieved from the Library of Congress, <www.loc.gov/item/2004669704/>.
Lee, Russell, photographer. Shepherd with his horse and dog on Gravelly Range, Madison County, Montana. United States Madison County Montana, 1942. Aug. Photograph. https://www.loc.gov/item/2017878800/.