There are many approaches to the building of a thesaurus and taxonomy. No absolutely right or absolutely wrong methods exist but there are some that are not as efficient as others. Some that lead to bias in the resulting work. Some that will not work with the data the client actually wants to surface on a website or data mine or tag. There are some very expensive approaches and some very practical approaches to doing the work. 

Ivory tower approaches are nice because they are not impeded in any way by the potential users. You just set up the world the way that you think it should be organized, fill in the details as you prefer and then try to sell the user groups on your ways of thinking. Sound far fetched? Take a look at the Rubricator created by Lenin as an outline of knowledge. It is the basis for practically all organization of information (knowledge) in the Former Soviet Union and the thinking of communism.

To really understand the way you think, you need to understand the way the information and learning are presented in our education systems, from preschool through advanced degrees. All information can be organized in different ways. The way that information is collected, tagged, and presented is the way that the readership will think about it. Card sorts and focus groups are a great way to get a view on how the experts would like to see the data organized, on a given day. They will have different answers and suggestions for organization three days later. The problem is that they do not hold up when you actually apply them to the data. Lenin’s outline is a wonderful idea, but when applied to people, like applying it to data it does not hold up. 

Why do you prefer going to Borders over Barnes and Noble? Look at the way they organize the books in the store. Online vendors like Amazon, Borders, Barnes and Noble have a different need in their approach. They need to allow several parameters to present the desired search result to book buyers. They have come up with taxonomies, with about 25 top terms for each. The terms are broad and quickly dwell down into the specifics of the piece. New, used, author, soft or hard cover, shipping restrictions, and other things covered by ONIX standards. They can use fielded (faceted) search for much of this just as LL Bean or eBay does. This data is presented from tightly controlled small lists or taxonomies and put into separate fields (elements) in a RDMS to allow intersection of the fields for the users to choose the parameters in search. 

Amazon has 14 categories on their web site at the top level. Click on them and you drill down to five or so more for each topic. I went on a field trip to some other bookstores to see where things were classified at those establishments. Some unusual placements to consider:

Borders Books Classifications
Food and Cooking
Psychology (which includes Magic and Erotica)
House and Garden (which includes transportation)
Pet & Nature
History and Politics
Music and Movies
Art and Architecture (which includes quilting and knitting)
Science Fiction and Fantasy
Mystery and thriller
Literature (I guess those other areas are not literature)
Young adult
Business and Computers (and wedding planning)
Bargain Books

They have a shorter list in airports but it includes electronics and games instead of some of the above topics.

Barnes and Noble Classifications
Crafts and Hobbies (including cookbooks and wines)
Health and diet (which has nursing and medicine)
Self Improvement (which includes relationships, family and children pregnancy and child birth, but not diet!)
Art (which includes gardening, and music, rock and roll)
Language (including education and home schooling)
B&N @ School
Barnes & Noble, Jr. (including Spanish language books)
Best sellers
History  (including law)
Philosophy (including religion and Christian aspirations)
Digital Photography (including programming and computers)
Reference (including wedding, photo albums and jigsaw puzzles)
Humor (including lots of games with a full subtitle for dice games)

Quite a difference between the two. I bet you feel more comfortable in one than the other because you understand the taxonomy they use to organize the store. The same is true of the grocery store and the hardware store — take a look at the signs.

There is a wide difference between the three big book approaches. I bet you feel more comfortable in one than the other because you understand the taxonomy they use to organize the store. The same is true of the grocery store and the hardware store, take a look at the signs, think about where you would store those items. Then think about how you present the information within your own organization. Is it easy and intuitive to find things? The taxonomy implied or actual may need review and augmentation to match the way you think of the content within your company. 

Access Innovations provides the tools and services to build relevant taxonomies.

Marjorie M.K. Hlava
President, Access Innovations

Related articles: