Site Index

	New England Chapter American Society of Indexers Past Presentations

Back to Fall 2005 Conference program page

Fall 2005 Conference Presentation

Taxonomies, Indexing, Search and Navigation:
Connecting the Dots 

Presented by Seth Earley

Seth Earley led an interactive discussion and slide presentation on the taxonomy derivation process and how it relates to indexing. He started off by soliciting from the audience ideas about what they wanted to learn about taxonomies. Audience suggestions included:
• Definition of knowledge management
• How to build a taxonomy for an intranet
• An explanation of the taxonomy creating process
• The best people to develop a taxonomy (and the trade-offs)
• Creating buy-in for taxonomies
• How freelancers get work with taxonomies
• How taxonomies are different from subject headings
• How to resolve multiple perspectives in context
• Where a controlled vocabulary fits into the process
• How taxonomies can improve search
• How to educate taxonomy users (deployment and socialization)

He then asked the audience of indexers what steps they follow when creating an index. The audience suggested the following:
1. Read the material.
2. Identify the audience.
3. Select content of significance.
4. Select terms for significant content, including synonyms.
5. Find and connect recurring themes/subjects.
6. Judge relationships among terms (and content).
7. Determine technology framework.

Seth then made the point that there is a lot in common between the indexing and the process of developing a taxonomy. The goals of a taxonomy, however, might be different from an index, because the taxonomy could be to serve either navigation or search, and what the goals is needs to be known from the start.

Issues in deriving a taxonomy, which must be address in advance, include the following:

1. Understand how the taxonomy is going to be used is important: metadata population, site or portal structure, navigational paths. Taxonomies must be connected to the navigation or search system to be valuable.
2. Unlike indexing, in which terms are attached to the content, taxonomies are an information framework. Content is hooked into the pre-existing taxonomy. Thus the taxonomy users must know how to apply the taxonomy language to the content.
3. Because the taxonomy is built in advance of the content, people don’t know what the content is. Those who create content, therefore, have to know the taxonomy; they must understand the mental model established by the taxonomist, so that they can coordinate their writing with the guidelines of the taxonomy.
4. After content has been attach to taxonomy terms, the users must understand these terms as well. They have to know what terms to look up; they must be more precise in their retrieval process. This is a major challenge to deployment.

Seth then delved into the complete multi-step process of creating a taxonomy.

1. Define the project and its boundaries (purpose, audience, taxonomy application method, user needs, and problems to be solved).
2. Determine the application framework, including inter-system relationships and updates and tools available.
3. Gather user behavior, both real and claimed ("imagined") behavior, such as from search logs and observation.
4. Survey content: collecting samples, mapping sources of information, looking for organizing principles, and looking at currency and ownership of content.
5. Perform a knowledge audit, which is really an exploration of context: how people solve problems, how people relate ideas to their needs, how different needs are related. User perspective is valuable in building a taxonomy that is both comprehensive and inclusive.
6. Review existing term sources, such as reference and academic sources and subject matter experts. The advantage of buying a taxonomy is that it’s a starting place, but the disadvantage is that revision—sometimes extensive—is almost always required. The distinction between what you buy and what you need is found within context: unless the context is exactly the same, revision is necessary.
7. Characterize audiences by how they approach their problems.
8. Organize terms by looking for themes. For this step, how terms are related is more valuable than the relationship itself. Notice that these themes may not be useful within an organization system, but they will help to define navigation.
9. Create “straw man” (an example that is meant to be criticized) and validate. This is where the early drafts are inspected as a reference point, evaluated and tested, and revised. As with all usability testing, the users are your best test subjects.
10. Create search and navigation scenarios. There are several tools and techniques, such as card sorting, that can help construct the application of all your analysis.
11. Iterate.

Seth finished up by presenting a case study of a sales resource center to demonstrate one example of the whole taxonomy building and application process. He presented the user interview questions and the overall project plan in MS Project. The example showed how taxonomic metadata is resolved with navigation.

Finally, Seth gave examples of how classification differs from a taxonomy, including different relationship examples.

---

Seth Early has been developing content and knowledge management systems for the past 12 years and has been in technology for over 20 years. He is an adjunct professor and Northeastern University, teaching knowledge management infrastructure and, as principal of Early & Associates, he currently assists enterprises in developing taxonomies for search, navigation, content management, document management, compliance and records management. Current clients include Textran, AIG, Fred Deimar, PTC, Parexel, Phillip Morris, and the Centers for Disease Control.
E-mail: seth@earley.com

---

Back to Fall 2005 Conference program page