Taxonomy is a system of organization, usually represented by a hierarchical list of controlled terms. If that’s still a little fuzzy, don’t worry, that’s why we wrote this blog. Let’s break it down.

The whole point of a DAM is to allow you and your colleagues to find what you need, when you need it. And a good taxonomy is the key, because well-organized and well-defined assets are easier to find and re-use. Organizing assets and setting up a well-understood structure gives users a point of reference that can be used universally by your colleagues. Better organization means greater ease of use of your existing assets, which means a better ROI on the materials you’ve already created.

Context is king

Taxonomies turn data into knowledge by adding context, which grants meaning to otherwise abstract terminology. Below are two examples that show how important context is.

It’s not hard to see how a search for a given asset can explode into a hundred unhelpful results. Taking the time to build a good taxonomy will ensure that you get the RIGHT results, not just a LOT of results. Instead of getting 400 records that you need to sift through to find something relevant, with a good taxonomy, you may get 14 assets, all of which are much closer to what you were looking for.

In librarian-language, creating knowledge from data by adding context is all about “ontological” work: essentially, the practice of defining the relationships that give your taxonomy its unique context. In layperson/everyday language, it means figuring out why your internal workplace jargon exists and using that knowledge to build a good taxonomy.

Building a good taxonomy starts by gathering feedback from the end-users of your DAM about how they access media to do their jobs. Integrating the user’s points of view will make your DAM a better system. To build a GOOD taxonomy, you’ll need to engage your users and all the folks who will be both upstream and downstream from the DAM. Talking through these processes with an eye towards building an inclusive metadata structure will help your asset management team get a more holistic understanding of all the steps that go into your creative processes.

Start the process of building your taxonomy with two questions:

• Who am I working with?
• What data am I working with?

Who am I working with?

Always, always, always start with your users. Explore their information-seeking behavior:

• How do they talk about the assets that they need?
• Do they prefer to start with a search or browse for content?
• Which seeking behaviors are based on necessity and which are based on preference?

Each type of user may have their own path (e.g., Producers, Photographers, Videographers, Editors, Graphic Designers, Executives, Client Users, etc.). Engage all your user-journey profiles while exploring information-seeking behavior.

What data am I working with?

Here are four questions to ask at the beginning of this process:

• What information do I have?
• What information do I want?
• Where is the information that I want?
• How do I get the information that I want?

Feedback is key

After addressing these questions and making any changes that came out of exploring the answers, re-engage your users for more feedback. Keeping your taxonomy healthy will involve soliciting ongoing feedback from engaged and thoughtful users with a stake in the project’s outcome.

A few feedback tips:

• Perform surveys that explore the terms your users prefer, or the ways that they’d prefer to tag assets.
• Have conversations one-on-one as much as possible.
• Look for statistics that support the need for a change of language or structure. Are there any specific searches that are returning too few or too many results?
• Have patience if your users want things changed; often people don’t know what they don’t like until they see it.

Taxonomy tips

We leave you with some additional food for thought...

Rule of Threes

Unless there’s a compelling reason, try not to go deeper than three levels in a hierarchy. Honestly, this is just a rule of thumb for trying to keep things easy to work with, conceptually.

AI is great, BUT...

If you start your process with AI as the first pass in tagging, it’s kind of like building a taxonomy in reverse. You often will wind up with a huge influx of data that will need to be quality-controlled and tweaked to “weight” the results for accuracy.

You’ll also need to do some work to train the AI to get it to recognize imagery accurately with the terms you’d prefer. This is an exercise similar to building a taxonomy from scratch, but you’re selecting your answers from a mass of data rather than generating terms from discussions with your users.