Leveraging Taxonomies to Improve Findability

by Zach Wahl

October 2011

Organizations today struggle with an increasingly challenging need to manage the mass of content they need to do their job. Only a decade ago, a relatively small set of business users were generating online content and even fewer were responsible for managing that content. With the democratization of content management, and more recently the ever-expanding social computing trends, anyone is a potential content publisher or manager.

As a result, organizations are bending under the weight of their own content. As more content is generated, it dilutes that which is truly critical to the business, creating issues with findability. One of the major drivers for building a taxonomy is to represent how people look for the information they need, and then categorize that information so it can be easily found. Taxonomies are most commonly associated with browsing and hierarchies, but taxonomies are also a key component of successful search experiences. The methods for using taxonomies to improve search are only beginning to be explored and developed.
This article discusses the ways taxonomies can improve search, navigation and overall findability. It details the various ways that users search for information and how taxonomy is an integral component of each. Specifically, it will focus on the way taxonomy enables the following key user activities:

• Searching for specific things in specific categories – Type and go or classic search.
• Browsing for general things in general categories – Navigate a taxonomy.
• Following links to related things that are located close together – Seek related content.
• Subscribing to information on pre-defined topics – Receive content automatically.

In each of these activities, effective taxonomy design and implementation is an integral component of success.

Searching
Some users know exactly what they are looking for—a specific document, picture, graphic, web site—and they simply want to retrieve this item so they can get on with their work. Information retrieval researchers call this known-item searching. The reason for filing cabinets and libraries is to make it possible to retrieve something when we know (or suspect) that it exists. The common type and go search interface is designed for people who know what they are looking for. They may get an even better user experience using some advanced search features such as author and title search, or qualifying a search by date. But they rarely use advanced search.
Behind the scenes a taxonomy can help type and go search in several ways.

• Adding synonyms to the words and phrases that a user types in. For example, adding International Business Machines and Big Blue to a search for IBM.
• Adding more specific or narrower terms to the search. For example, adding Training and Job Openings to a search for Careers.
• Adding more general or broader terms to the search. For example, adding Arms Agreements to a search for START II.
• Adding related categories to the query. For example, adding Flash Memory Cards, Camera Batteries, and Camera Cases to a search for Digital Cameras.

Of course, searchers will not want this help all the time. The information retrieval application may let them control such additions by separating the additional results from those that match just the original search.

Browsing
Some users don’t know exactly what they are looking for. In 1982, information science researchers Belkin, Oddy, and Brooks named this condition an anomalous state of knowledge or ASK. Some users with a less articulated search strategy find it useful to browse a set of general categories or navigate a taxonomy like the Yahoo Directory or the DMOZ Open Directory Project (dmoz.org) shown in Figure 1.
Figure 1 - DMOZ Directory
Searching for information should be like shopping online—you start with a high-level set of categories like the Yahoo Directory, or you type in a few words and then navigate the results of that search.

Following Links
Scholars and researchers have used footnotes and citations for hundreds of years. Once a key document is found, chances are that the sources cited will be worthwhile to review, the sources in the sources should be relevant as well, and so on. Not only is it the convention to reference the sources of an explicit quotation or an implicit argument, but tracking citations has been shown to be a useful way to study the evolution and history of ideas.
Taxonomies are important in citation indexing because building the indices relies on authority files of author names, journal titles, publishers and article titles to make the matching of bibliographic citations more accurate. The Dublin Core Metadata Initiative (dublincore.org) broadened the notion of authorship to include the creation of all types of content objects such as art works, business records, etc. This broader notion of creator thus includes artists and artisans, company officers and employees, etc. Social network analysis is based on the ability to identify people and the relationships between them. So taxonomy is important for effective social networking, because authoritative naming makes it possible to identify the right person you mean to link to or tag.
Taxonomies are important in helping to identify what related information should be linked to in a web page, especially when that information frequently changes such as government rules and regulations, or when new information is routinely published such as press releases and news stories. Errore. L'origine riferimento non è stata trovata.Figure 2 shows the U.S. Environmental Protection Agency’s climate change homepage (epa.gov/climatechange) news links. This sort of sidebar of news links is very common, and is one of the ways that taxonomies can be used to help drive dynamic web sites.
Figure 2 - EPA Climate Change Webpage News Links
Subscribing
Another way to find things is to come across them in newspapers or magazines, to hear them on the radio, to see them on television, to receive pointers in messages by signing up to old-fashioned mailing lists, and more and more to read them on the web. Many people still subscribe to a few newspapers or magazines. Today we also subscribe to multiple mailing lists for announcements, listservs and blogs, follow people on Twitter and Facebook, and use other social media. We may also be registered users of a growing number of web portals for work, shopping, and other personal interests. The scholarly journal is nearly 300 years old; and today academics and professionals continue to follow their field or specialty in scholarly and trade periodicals. With the growth of information outlets, there is a greater desire to receive more specific information related to our specific duties and interests.
The most common syndication method today is RSS, a service available on many websites. RSS stands for Really Simple Syndication and provides the service to be alerted to newly published or updated web content. Taxonomy is important in providing the categories to refine and target what is included in a specific RSS feed. The EPA offers RSS feeds based on the issuing agency and a set of approximately 25 broad topics such as Air and Water. The New York Times offers RSS feeds based on the section of the newspaper sub-divided by broad topics. With more and more web-accessible content, more automated alert and filtering services will be coming in the future.
Conclusion
Though the concept of taxonomies continues to be relatively unknown to a majority of content consumers, virtually anyone acting to find information will leverage taxonomies. As such, user-centered taxonomy design is an increasingly important consideration when capturing, managing, or presenting content. Effective taxonomy design will enable multiple avenues for findability, resulting in improved information use and access.