Sunday, June 24, 2012

Enterprise Taxonomies vs. Traditional Taxonomies

A book that I have been reading (Structures for Organizing Knowledge: Exploring Taxonomies, Ontologies, and Other Schemas, by June Abbas, 2010) got me thinking about the comparison between corporate/enterprise taxonomies and other “traditional taxonomies”. I found it intriguing that Abbas presents corporate or “professional” taxonomies in the same chapter on personal information structures. Thus, a corporate taxonomy could more aptly be an extension of a personal knowledge organization system, rather than the customization of standard taxonomy or controlled vocabulary.  So, how are corporate taxonomies or enterprise taxonomies (corporate taxonomies that are specifically for use enterprise-wide) different from traditional (library science type) taxonomies or thesauri?

There are, in fact, multiple ways in which a corporate or enterprise taxonomy differs from the traditional taxonomies or controlled vocabularies used in libraries or in particular subject disciplines. Enterprise taxonomies in particular are:

1.      Relatively small in size
2.      Multifaceted
3.      Customized to an enterprise’s content
4.      Customized to an enterprise’s users
5.      Relatively informal

An enterprise taxonomy tends to be relatively small in size with respect to the number of terms and depth of term levels. The size will depend largely on the complexity of an enterprise’s business (number of lines of business, for example), but the range of 1000-2000 terms in an taxonomy for an enterprise that has single line of business is typical. An organization may certainly supplement  this enterprise taxonomy with additional subject-specialized controlled vocabularies, particularly in the areas of research & development or product catalogs.

Faceted Nature
An enterprise taxonomy deals with a variety of content which is differentiated in more than one way, not just by subject matter. Content is typically organized and searched not merely for what it is “about” but also what its purpose is, what its source is, what type of content it is, and perhaps also for what market or customer type it is relevant. Thus, an enterprise taxonomy is usually organized into several facets to support faceted search or faceted browse (see my April 2012 post), which include: document type, file format, department or functional area, line of business or product/service category, geographical region, and market segment, in addition to a topical facet.

Content Customized
A corporate or enterprise taxonomy should be highly customized to an enterprise’s own unique content. While two companies in the same industry may have nearly identical products and services, their customer or member base could vary slightly, and they probably do not have identical organizational structures, procedures, and workflows. Thus, no two companies or organizations would have identical content. Organizations also differ in the quantity of different kinds of content they own and in the importance they assign to different types of content.

User Customized
Just as important as content-customization is user-customization. Corporate or enterprise taxonomies are designed to help an organization’s users (employees, and often also partners and customers) find content. Users include both those who upload/publish content to the intranet or content management system, often manually tagging it, and users who are looking for content. These are sometimes the same people and sometimes not. Also in consideration of the users, there may be a workflow or business rule aspect that is taken into consideration. Thus, the process of designing an enterprise or corporate taxonomy involves gathering input from users, via interviews and workshops. For this reason, the author Abbas has combined corporate taxonomies into the same chapter as personal taxonomies, because they are both highly user-centered.

Traditional discipline taxonomies (such as for living organisms), thesauri, book cataloging and classification systems follow industry standards for their design and construction, which can be quite rigid and formal. For general-purpose controlled vocabularies, there are the ANSI/NISO Z39.19  guidelines and ISO 25964-1 standard (see my March 2012 post), which allow more flexibility than library cataloging rules. The design of corporate or enterprise taxonomies should adhere to ANSI/NISO or ISO standards at a high level, but in practice, other practicalities and user needs and expectations should take precedence over a strict following of every detail of the standards.