Before You Implement a DAM — Consider Your Taxonomy!

by Heather Hedden

Any digital asset management (DAM) system requires taxonomy management; but, how to create an effective taxonomy, implement it, and use it properly is not a simple task.

Issues in taxonomy definitions

The word taxonomy has varying meanings. A taxonomy is a controlled arrangement of terms used for tagging and then retrieving content. It can be more narrowly defined as a hierarchical arrangement of terms for topics, where broad categories have more specific categories under each. A taxonomy can also be more broadly defined as a controlled vocabulary, or set of controlled vocabularies, in any structure. Its purpose is to provide a restricted list of terms for tagging content, rather than coming up with or randomly generating new key words. It’s best to consider a broad meaning of taxonomy, because a DAM or other kind of content management system will support both hierarchical and flat-list taxonomies.

The word taxonomy may also be used inconsistently to refer to one or more than one controlled vocabulary. It may refer to the set of terms for a single purpose and scope, such as a taxonomy of geographic places or person types, or a taxonomy can refer to the entire set of controlled vocabularies used in the same implementation for the tagging and retrieval of the same set of digital assets, especially when used in combination as filters. So, people might refer to the taxonomy (singular) or taxonomies (plural) in their DAM, and they mean the same thing.

Taxonomy value and benefits

The primary benefit of a taxonomy is not merely its hierarchical structure but also that it’s a controlled vocabulary, so that people (both those that tag and those that search/retrieve) use the same words to describe the same thing, whether they are true synonyms (e.g. Cars and Automobiles) or something similar enough to be grouped together (e.g. Food Preparation and Cooking, Fundraiser and Charity Event, or Seashore and Beach).

If the taxonomy is small enough to be browsed, both the taggers and the end-users select terms from the same controlled list. If the taxonomy is too large to be browsed and must be searched, then “synonyms” (alternative labels or variants) are included in the taxonomy to redirect the users who search on these variants to the approved term.

The benefits of a taxonomy over keyword search is that users don’t miss retrieving assets because of use of a different synonym, and users don’t retrieve undesired assets because they were described with words that have multiple meanings.

The benefit of a taxonomy structure is to guide end-users to find the appropriate taxonomy term and linked assets, if users do not know how to describe what they want or are not certain what they want and prefer to browse the collection. A hierarchical structure guides users from broad categories to more specific topics.

A faceted structure guides users to assets in a slightly different way, by enabling users to browse and select combinations of descriptions or attributes, such as location, person type, topic, event/occasion, etc., as filters.

Taxonomy practices and risks

Taxonomies can be very useful, but they must be:

  • Designed well,
  • Customized for the particular set of assets and users,
  • Utilized properly in indexing/tagging, and
  • Maintained

Taxonomy terms and structure/relationships should be designed, to the extent possible, in accordance with accepted best practices or standards. The most relevant standards are ISO 259654-1: Thesauri for information retrieval, and ANSI/NISO Z.39.19-2005 Guidelines for the Construction, Format, and Management of Monolingual Controlled Vocabularies. What is particularly helpful for taxonomies in these standards is term name format and the proper use of the hierarchical relations.

A taxonomy should reflect the precise scope, depth, and breadth of the content (set of digital assets) it is intended to tag and retrieve. Thus, usually taxonomies are custom-created. It may be appropriate to license a third-party taxonomy (small controlled vocabulary), for one or two of the facets of a faceted taxonomy, such as geographic location or industry, if the content scope is sufficiently broad. Even then, modification of the licensed taxonomy is probably needed.

Most importantly, a taxonomy must be designed to reflect the needs and preferences of its users, such as technical wording for a technical audience and common wording for the public.

A taxonomy also needs to be implemented properly in indexing/tagging the assets. For example, it’s not helpful to users if assets are tagged with too many or too few taxonomy terms. And, if a faceted taxonomy is being used, assets should be tagged with a term from each facet. Therefore, policy for tagging is needs to be written, and those who do the tagging should be trained on the policy.

Finally, a taxonomy needs to be updated and maintained to reflect the growth and change in the collection of digital assets. Taxonomy work is never done.

Resources:

Hedden,  Heather. “Taxonomies and Metadata for Digital Asset Management.” (pdf) Journal of Digital Media Management, Vol. 6, No. 4, Summer 2018. pp. 380-387.

Taxonomy creation workshops and courses, Hedden Information Management

Heather Hedden is a professional taxonomist, a taxonomy consultant running  Hedden Information Management, and the author of both a book and a blog called The Accidental Taxonomist http://accidental-taxonomist.blogspot.com/

She will be in Amsterdam on on 21 March, 2019, giving a 2-hour Introduction to Taxonomies Workshop, at the Vogin-IP-Lezing conference “zoeken & vinden.”

This article, written for our website, is Copyright © 2018 by Heather Hedden.

Taxonomy Licensing by Heather Hedden

Heather Hedden, of Accidental Taxonomist book and blog fame, published a new article on licensing taxonomies.  She discusses a surprising variety of unexpected issues you can face without a thoughtful, customized taxonomy and she provides resources where you can find open-source, or license, taxonomies and controlled vocabularies, including the Taxonomy Warehouse, and the University of Basel Library, the Basel Register of Thesauri, Ontologies & Classifications (BARTOC). .

There is also a multiple-choice survey, which results she will post.