Who Should Create Taxonomies?

More and more organizations of various types and sizes are recognizing the benefits of information/content taxonomies, to make it easier to more accurately and quickly find information, be recommended information, and be able to formulate complex queries of data.  

In many cases, however, where taxonomies are not central to the product/service of a company (such as e-commerce retail or information publishing) or function of an organization (such as research), the task of creating and maintaining a taxonomy is not big enough to justify hiring a professional taxonomist. Creating a taxonomy is a temporary project, and then updating it is often a part-time task, which could even be shared among several people.

Taxonomy creation should not be underestimated, however. It may appear easy to create a taxonomy, but it is not easy to create a good taxonomy. If a taxonomy is not well-designed it cannot serve its purpose well. You may as well rely on a search engine alone than try to utilize a bad taxonomy.

Not creating the taxonomy yourself

Some approaches to developing a taxonomy without a dedicated taxonomist include using existing taxonomies, creating a taxonomy by term extraction, or hiring a consultant.

Reusing existing taxonomies

To serve its purpose best, a taxonomy should be custom-created to serve its content, users, and system. An existing external taxonomy is usually not adequate. It may be suitable for limited scope of a geographic taxonomy, industrial classification, a list of organization names, a list of languages. More information about licensing taxonomies is in my blog post “Taxonomy Licensing”  Even when using an existing taxonomy, there is still work to edit and adapt the external taxonomy, which requires taxonomy expertise

Creating a taxonomy by automatically extracting terms from content

Software, including some taxonomy management software, such as PoolParty, can extract candidate taxonomy terms from a body of content (documents or web pages) that is intended to be tagged with the taxonomy. This is an effective method to enhance a taxonomy, to add missing concepts and alternative labels (synonyms). However, this is not a practical way to start creating a taxonomy, which requires a logical structure. Taxonomy-creation expertise is still needed.

Hiring a taxonomy consulting or temporary contractor

This is a good idea. A consultant or contractor will provide a combination of guidance and actual taxonomy building, although a consultant tends to provide more guidance, and a contractor tends to do more taxonomy building. A contractor requires a certain time commitment, such as 3-6 months full-time, whereas there is lots of flexibility in engaging a consultant. After the consultant or contractor is finished, though, someone needs to maintain and update the taxonomy to the same specifications.

When a taxonomy is not very large, it may be more efficient and cost-effective to create it from scratch oneself without reusing an existing taxonomy or relying on a consultant or contractor, although getting a consultant to at least review the taxonomy might still be a good idea.

Taxonomy management as part of a role

What is much more common for an organization than to have a taxonomist is to have one or more positions where taxonomy management is part of the job description. Searches on web job boards return hundreds of job opening with “taxonomy” in the job description, whereas only a small fraction of them have taxonomy or taxonomist in the job title. Common job titles include:  Content Designer, Content Manager, Content Strategist, Data Architect, Data Catalog…, Data Strategist, Digital Asset Manager, Digital Content…, Digital Librarian, Information Architect, Information Scientist, Knowledge Engineer, Knowledge Management…, Metadata Specialist, Product Manager, SharePoint Developer, Solutions Architect, etc. There are also positions more centered in marketing and in web development.

Often, though, the need for a taxonomy emerges at a time when a new position is not created, so an existing employee must take on the task. This common scenario is behind the title of my book and this blog, The Accidental Taxonomist. Those that take on taxonomy work may come from a wide variety of roles or departments including marketing for a website taxonomy, IT or human resources for an intranet taxonomy, IT for content/document management systems administration, and technical documentation/publishing. Knowledge management and metadata/data management are also good candidate roles for taxonomy management.

In situations where the taxonomy is used to manage and retrieve content in specialized subject areas, subject matter experts may also be involved in taxonomy creation, at least for the parts of the taxonomy that correspond to their expertise. 

Not having sufficient taxonomy skills

In either case, whether taxonomy management was originally part of the job description or not, people who assume partial taxonomy responsibilities often do not have the skills. This is usually the case when a taxonomy project first arises. Even when someone is newly hired, successful applicants may not to meet all job description duties, such as taxonomy experience, especially if the skill is only a minor part of the job.

Related job skills may make it easier to created taxonomies, but without experience or training, one cannot simply create a good taxonomy. Related skills tend to be in the area of library/information science, indexing, information architecture, digital asset management, content management, records management, and possibly product management.

Librarians tend to have training in cataloging and classification, sometimes in thesaurus creation, and less likely in taxonomy creation. Taxonomies resemble classification schemes, but function differently, so it would be a mistake to model a taxonomy as a classification scheme. See my blog post "Classification Systems vs. Taxonomies." I had taught a continuing education course on taxonomies through a graduate school of library and information science for years, since MLIS graduates had not learned taxonomies as part of their degree program.

Information architects know how to organize information in a web user interface well, so they may have a good sense on how to structure a taxonomy at a high level. However, there are details and nuances of a large taxonomy, such as the development of synonyms/alternative labels, with which they may not have experience. Also, a taxonomy should not be confused with a navigation scheme, as explained my blog post "Navigation Schemes vs. Taxonomies."

Digital asset managers, content managers, and product managers know about the metadata management for their content, and taxonomies usually fit into the larger metadata scheme. However, their experience with taxonomy creation is usually limited to a subject area and the context and constraints of the system in which they are working. So, the very basic taxonomy skills that they develop may not be transferable to another system or another subject domain.

Subject matter or domain experts, including product managers, often play an important role in taxonomy development. From my experience in working with subject matter experts, though, they often tend to design more of a classification scheme for their domain and create taxonomy concepts that are too granular to be practical for end-using search and retrieval.

Where to learn taxonomy skills

There are many continuing education options to learn taxonomy creation, some through library/information science schools, some through professional associations, and some through commercial conference and training programs. I have been providing taxonomy training since 2007, through online courses, conference workshops, and corporate workshops, both in-person and virtual. I have been impressed with the diversity of backgrounds, job roles, organization types, and global locations of the workshop participants over the years.

I teach various online and in-person workshops. The latest information is on my website page Courses and Workshops.

Upcoming in fall 2021:

"Knowledge Engineering of Taxonomies, Thesauri, and Ontologies"
SEMANTiCS conference, Amsterdam, Monday September 6, 9:00 am - 12:30 Central European Time. Available both online and in-person.

"Taxonomy and Metadata Design"
Monday-Tuesday, October 11-12, 8:00am - 12:00pm EDT (14:00 - 18:00 CEST) each day (8 hours over two days)
Through Technology Transfer, Rome (with the availability simultaneous interpretation into Italian).