First Aid for the Accidental Taxonomy

Many successful information systems utilized taxomies and metadata, but finding taxonomists to support this work usually happens by accident. Taxonomy design and development is a specialized skill – maybe even a talent. A large organization may employ information architects, SharePoint architects, content managers, and corporate librarians, but these people most likely lack strong taxonomy experience. Although the closest matching formal education for taxonomy work is a masters in library and information science, many corporate librarians specialize in research (such as in business intelligence) and may only know about classification and organization of information based on a past library school course taken. Information architects’ experience with taxonomies may be limited to small taxonomies that fit within the limits of menu labels.

When an organization decides it needs an enterprise taxonomy or needs to leverage and redesign existing taxonomies, then any of these aforementioned types of employees are often pressed into working as taxonomists, without prior experience.   This is what I refer to as the “accidental taxonomist,” as in the title of my recent book (see Heather Hedden, The Accidental Taxonomist, Information Today Inc., 2010 ).

While reading my book is a good idea for anyone who becomes an accidental taxonomist, a book alone cannot teach all the needed skills. Designing and building taxonomies is a process that is fraught with decision-making. If a taxonomist or taxonomy manager is not a defined position within an organization, the “accidental taxonomists” who temporarily assume this role, no matter how skilled, still have their regular jobs to do and may not be able to devote the needed time for the taxonomy.

Starting with a good taxonomy foundation will make it easier to maintain the taxonomy. It will save money, time and resources to get some outside help, especially during the initial stage of taxonomy development, which requires the greatest investment of hours.

How can a taxonomy consultant help?

  • How should terms be assigned to facets
  • Should hierarchies be more deep or more broad
  • Is a complex hierarchy needed or will simpler arrangements work
  • Should taxonomy term labels be complex or simple
  • What governance is needed for longterm management of the taxonomy
  • Who should be on the governance team, and what training is needed

Related to different levels of experience there is also a distinction between explicit knowledge, which may be explained in a book, and tacit knowledge, which is gained through expertise and is more difficult to explain or document. Taxonomists are trained to follow the industry standard guidelines, such as ANSI/NISO Z39.19-2005 Guidelines for the Construction, Format, and Management of Monolingual Controlled Vocabularies. But these are just “guidelines,” and in practical applications the guidelines may need to be modified slightly, such as when there are significant restraints on the taxonomy design. Knowing where and when it is appropriate to bend the rules and when it is not, is a part of tacit knowledge. Having the right knowledge, however, does not necessarily mean the taxonomy work gets done.

Even within the narrower area of taxonomy expertise, it often helps to discuss and work out issues among multiple people who have an understanding of taxonomies Taxonomy work in a large organization can be a team effort. It requires different skills and perspectives to serve all its goals. In addition to the lead taxonomist with an information science background, other people needed include information architects and user experience professionals to ensure that the taxonomy fits well into the user interface and is easy to use, subject matter experts as authorities on the terminology, and IT professionals for the technical implementation of the taxonomy.

If you are the sole taxonomist in your organization, you may want to consult with other, outside taxonomists, such as through online discussion groups, to bounce your ideas off them and get additional feedback based on their varied expertise. It’s hard to work alone without support.

If you are at SLA Annual Conference in New Orleans on June 16, come hear my talk : “Taxonomy Made Easy: An Introduction to Taxonomies for the Accidental Taxonomist.” SLA members are mostly corporate librarians, who are likely candidates to become accidental taxonomists. I’ll help you develop your own taxonomy skills and also identify where you might need to talk to your management about consulting with others skilled in taxonomies. First aid for the accidental taxonomist is always available!

~ Submitted by Heather Hedden

Enhanced by Zemanta

How to Teach Taxonomies

On several occasions   Heather Hedden, one of our authors,  gives full-day workshops on how to create taxonomies and other controlled vocabularies, and it’s interesting how different issues or problems arise in different sessions. It may be just one individual who has difficulty grasping a concept and asks lots of questions, but in answering the questions, it turns out that others have similar questions. It then becomes apparent that some principles, while simple on the general abstract level, can get muddled when we start looking at specific examples.

Here are Heather’s strategies that she’s  learned to help train new taxonomists in building taxonomies:

1. Use different methods of explanation to serve different backgrounds and mindsets

In one session I gave, an exercise on creating correct polyhierarchies had some people perplexed. The exercise was to propose two or more broader terms for “Paint Brushes.” Had I asked the participants to suggest just a single broader term, I probably would have gotten mostly correct answers, such as “Painting tools” or “Brushes.” But when having to propose two broader terms at once, many taxonomy students had gotten off-track and proposed a pair such as “Artists tools” and “Contractors’ tools.” This, of course, is not correct, since not all paint brushes are used by artists, and not all paint brushes are used by contractors. In one session, several minutes of attempted explanation to one individual were insufficient.

For the trainees who are more mathematical, an analogy with Boolean logic might make more sense. In such a case, we can say that the polyhierarchy represents the Boolean AND, rather than OR. The narrower concept must always belong to both Broader Term A AND Broader Term B, and thus the narrower term in a polyhierarchy represents the intersection or union of its two broader terms. For others with a reference searching or indexing background, however, it is necessary to explain the search and retrieval implications of the hierarchy. For example, “Paint brushes” is a term indexed to documents on all kinds of paint brushes, including those for commercial exterior painting. Thus, it would be incorrect to retrieve the latter documents with the broader term for “Artists’ tools.”

2. Teach the standards in the most practical sequence

At a recent corporate training I gave, the topic that was most challenging for the participants was associative relationships. The specific issue was what kinds of term pairs could legitimately use this kind of relationship. Listing all the possible types of term pairs for associative relationships as described in the ANSI-NISO Z39.19 standard (13 of them!) can add more confusion than needed. Training participants wanted to refer to this list when completing an exercise on proposing related terms. As a result, they proposed some rather unconventional related terms, that, while legitimate, were not entirely practical.

Teaching principles from the ANSI-NISO standard, I realized, should not necessarily rely on the same sequence or emphasis as written in the standard. It is probably better to present the list of 13 associative relationship types at the conclusion of an explanation on the associative relationship, rather than at the beginning. Similarly, I found that the ANSI-NISO standard’s order of presenting relationships between terms belonging to the same hierarchies followed by term terms belonging to different hierarchies may also not be best, since relationships between terms belonging to different hierarchies are more common.

3. Simplify demonstration exercises

Setting up a practice card-sorting exercise for classifying things can be a part of training session, but the card-sorting exercise should be modified. I had attempted once to create a number of cards (at least 25) considered desirable for a card-sorting exercise, only to find that the training participants only want to spend a few minutes on it, rather than the more intense 10-20 minutes that actual card-sorting participants would be expected to devote. Thus, a demonstration card-sort should be a much small set, along with the clear statement that an actual card-sorting exercise would be larger.

4. Revise sample exercises

Finally, the exercises included in training workshops may need to be tweaked after they are tried out. For example, asking for suggested broader terms for the term “Financial Management” obtained such varied and questionable results, that I had to remove that example and replace it with another. What I had in mind as broader terms were quite simple “Finance” and “Management,” but my audience was looking for other, more specific meanings of the term. Examples should not be ambiguous.


Even if you are not a professional taxonomy trainer, a lot of training is needed in the taxonomy field, and a lot of people learn how to create taxonomies on the job. Thus, if you are a taxonomist, you may find yourself required to teach taxonomy principles to others. Although some of it may come naturally to you and those you teach, other taxonomy principles will be more difficult to teach, and creative strategies are needed.

– Heather Hedden

Heather can be contacted for more information about her workshops and upcoming presentations at  Hedden Information Management.

Book Review: Organising Knowledge by Patrick Lambe

Although the interest in and applications of taxonomies has grown in recent years, there are still not many books on the subject. Most of the information on taxonomies currently resides in online discussion group archives, blogs, wikis, conference presentations, white papers and reports (the latter at quite a premium price), but not much yet in easily accessible books. A search on on “taxonomies” yields numerous books of specific taxonomies, but very few on the art of creating taxonomies in general. Even the “books” page on the Taxonomy Community of Practice Wikispace lists mostly books on information architecture, a classic book on classification theory, chapters of books on broader topics, and high-priced research reports. There is just one book listed with a focus on taxonomies: Organising Knowledge: Taxonomies, Knowledge and Organisational Effectiveness by Patrick Lambe (Oxford, England: Chandos Publishing, 2007)

Indeed, as its title and subtitle suggest, taxonomies are presented within a broader view of how knowledge is organized. The book is neither a simple “how to” book, nor a scholarly treatment of the subject, but in fact combines both: practical advice on how to create taxonomies along with thoroughness in covering the field of knowledge organization and analysis of various ideas and previous literature on the subject, with many footnotes and a lengthy bibliography.

The author, Patrick Lambe, is a Singapore-based consultant in the field of knowledge management who can base his ideas on his own business experience. Yet Lambe also has the academic credentials of an information scientist, a Master’s degree in Information Studies and Librarianship and experience teaching as an adjunct professor. Thus, he aptly bridges both sides of taxonomies, the traditional library science side and the newer corporate knowledge management side, although it is the latter that is the subject of this book. What I appreciate in this book is that Lambe writes based on both his research and his experience, and based on these he has developed a number of his own ideas.

While common definitions of taxonomies often limit them to hierarchies, Lambe prefers a broader definition. The forms of taxonomies that Lambe presents, along with a detailed explanation for each, are: lists, trees, hierarchies, polyhierarchies, matrices, facets, and system maps. Stretching the definition and boundaries of what taxonomies are and can do is a central theme of Organising Knowledge. Lambe states: “Taken together, it becomes clear that taxonomy work holds a wider range of application and use than simply as a tool of information retrieval.” (p. 95) .

Organising Knowledge presents a number of real world examples, scenarios, and case studies of the application of taxonomies in their broadest sense. These include implementations by the U.S. Department of Homeland Security, Unilever, and Club Med. These examples illustrate the wide range of uses for taxonomies. Among business activities, Lambe says that taxonomies can support the areas of risk recognition and response, cost control, customer and market management, and innovation.

Lambe does not simply describe taxonomies and their use. In this in-depth book he discusses their varied roles, how they are understood, and trends in their implementation. He describes how different kinds of taxonomies can either (1) structure and organize (both things and processes), (2) establish common ground, (3) span boundaries between groups, (4) help in sense-making, or (5) aid in the discovery of risk and opportunity.

Several later chapters turn to the practical steps of preparing, designing, and implementing a taxonomy project. Lambe breaks out the process into ten steps, the first six of which are all still part of the preparation stage. Among the topics presented in the preparation phase are taking technology into consideration and communicating well with the taxonomy sponsor and stakeholders. While it is appreciated that technology/computer systems are mentioned, I would have liked to learn more about this. It becomes quite evident that different situations require different approaches and different kinds of taxonomies, the different kinds of taxonomies that Lambe describes earlier in the book. My only point of disagreement here is the continual distinction between tree taxonomies and faceted taxonomies, since taxonomies often exhibit both characteristics at the same time.

The book is well written and relatively easy to follow, but it is not a “light” read. It has a number of helpful tables and diagrams. Particularly useful is the table (two and half pages long) comparing the uses and issues for each of the seven forms of taxonomies: lists, trees, hierarchies, polyhierarchies, matrices, facets, and system maps.

I highly recommend this book of great breadth and depth to anyone who works on taxonomies or is interested in working on taxonomies. The intended audience of the book is indeed limited to knowledge management and taxonomy professionals. Even those with considerable experience working in taxonomies will find this book informative and enlightening.

– Heather Hedden

This review is based on a longer book review written by Heather Hedden and published in Key Words, the Bulletin of the American Society for Indexing, Vol. 15, No. 4, October-December 2007, pp. 130-132.