Taxonomies are used to organize information to improve discoverability and control. This could be website navigation, folder structures, or metadata groups. The right taxonomy depends on many factors, and I often used Dave Snowden´s Cynefin framework to discuss this with clients.
Known, knowable, complex, or chaotic information?
-
Known: You then know your information and you can rely on structures that does not often change. You can then rely on a hierarchical arrangement of terms that people agree upon.
-
Knowable: People may then have different ways to find the same same information.You may still rely on hierarchical arrangements, but improve this with synonyms, best bets, and facets allowing multiple ways to find the same information.
-
Complex: For complex areas, rely more on standardized metadata element and values supported by social tagging. If a new social tag is used a lot, include it then in the standardized metadata model.
-
Chaos: Rely on user-driven social tagging to improve discoverability and control of information. Other frameworks may be out-of-date before they are implemented, or only work for some user groups.
Most organizations will have information types that cover all four areas, e.g. Financial and HR information will be known, R&D information will be complex. A taxonomy will therefore have to be fit-for-purpose and often requires constant change.
Types of taxonomies
Here is a list of your common options from Patrick Lambe, author of Organizing Knowledge:
-
Lists – simple collection of related things like business areas, legal entities, etc.
-
Trees – break a list into sub-categories, and useful when lists are too long into sub-categories
-
Hierarchies – similar to trees, but the children inhere the same attributes as the parents.
-
Poly-hierarchies – similar to hierarchies, but allow for “virtual linking” of children that belong to several parents.
-
Facets – multi-dimensional taxonomies that are comprised of multiple metadata elements, e.g. wine classified based on price, region, grape type, etc.
-
System maps – a visual representation of a domain of knowledge, e.g. London Tube Map
How to develop a taxonomy
Here is a recommended approach:
-
Identify stakeholders
-
Define purpose
-
Determine approach
-
Collect information
-
Develop taxonomy
-
Pilot taxonomy
-
Deploy taxonomy
-
Gather feedback
Taxonomy success factors
Patrick Lambe suggests the following success factors:
-
Intuitive – users can determine where to find content by looking at the top level and do not get lost in the structure.
-
Unambiguous – users do not have multiple places where a particular record might be found or placed.
-
Hospitable – the scheme covers all current content and expected future content without massively overhauling it.
-
Consistent – The scheme is consistent at different levels or in different areas.
-
Relevant – The scheme is representative of the way the organisation works, including both processes and terminology.
-
Parsimonious – there are no categories that are unused and no categories of undifferentiated stuff, such as “General”, “Miscellaneous”, “Documents”, and the like.
-
Meaningful – categories and terms agree with their content and match general usage.
-
Durable – The scheme does not require frequent or radical updates.
-
Balanced – Categories, subcategories, and content are all dispersed evenly across the scheme.
I learned at a Web 2.0 conference many years ago, that the key to success when developing taxonomies that will be used by people:
“Don´t make me read, don´t make me think”….
Feel free to contact us if you need help establishing an architecture foundation for your organization with the required taxonomies.
You may find the following blogposts of interest:
