There are five kinds of taxonomy strategies available to you depending on the different situations your business finds itself in.
Structure & Organize
The Structure & Organize taxonomy project is for systems and content that are already very stable, well documented, and broadly understood by your user communities. There is already a lot of consistency in how people are describing and organizing the content. This takes in the majority of your business records.
An example of a Structure and Organize project is the creation of a taxonomy for a technical education college’s curriculum development system. All of the curriculum documents, topics of study and levels of study are very stable and widely understood. Now a single online system with a curriculum development workflow needs a taxonomy to enable different people in the system to pull out and filter the content according to their needs.
This kind of project is the easiest of all taxonomy types to develop, it is simply a matter of documenting the language in use, using a tree structure if the system is simple and with limited content, or identifying the most useful facets if there is a lot of content or diverse user groups, and then tidying up any minor inconsistencies.
Here you have users who are conducting the same kind of functional role, but are separated physically and work in different teams from each other. As a result, they have evolved different and inconsistent working practices, and they have developed different ways of describing and organizing their information.
Building Common Ground taxonomy projects arise when you want to get efficiencies from more consistent work practices across teams in the same function, and better information sharing, learning and knowledge reuse across the same function. Your task is to try to negotiate a taxonomy that all the different teams can agree to.
If the information environment is relatively simple, and differences are not too extreme, then a tree structure might suffice. But for more complex information environments, and where there are significant differences in perspective and in how people want to organize and approach their information, then facets are going to be the best solution.
For example, in quality management, some teams organize there documents by process, some organize by job role, and some organize by document type. Rather than forcing two thirds of the teams to follow a standard dictated by just one, thereby creating great disruption to established work patterns, you can solve this problem by developing a facet for each perspective and ensuring that all documents are tagged for each facet.
Your task is first to understand and document the differing perspectives and then negotiate a shared standard that everybody can use, respecting the diversity as far as possible.
Dealing with specialized knowledge and expertise, the problem for your organization is how to ensure that knowledge and information flows effectively between these specialized work communities, and between specialized experts and non-specialists to support effective coordination of tasks and the transfer of learning and ideas.
An example of a Boundary Spanning project might be to create a taxonomy that allows a customer service team to share knowledge and information with an engineering team. This might be to ensure that customer service does not give inaccurate information to customers, and to ensure that engineering understands the problems that customers want to be solved.
Facets are almost always necessary in a Boundary Spanning project, because facets are the best way to project the same content through different perspective lenses.
Where multiple knowledge communities are sharing the same information system, the taxonomy system needs to be more complex, and here the facets will need to be backed up with thesaurus and possibly an ontology.
Here your role is usually to start documenting the language and organizing habits of the different user communities, identify relevant facets, and start mapping relationships between the language used by one community and the languages used by other communities, when referring to the same things.
As you get into more complex areas of content and business, your taxonomy shifts from becoming a system that describes what is already known, towards a framework for interpreting what is not completely known. In this environment, rules do not help, and applying expertise or analysis does not work. To address this environment, you need to look for patterns, try to resource the patterns that look good, and avoid the patterns that look bad. You resource multiple small bets rather than a single major strategy because the situation is changing all the time, and you don’t understand enough about it to make a major commitment.
This means it becomes much simpler and lighter in structure. Matrices, system maps and other ways of visualizing and sorting information into broad patterns are much more useful than detailed trees or ontologies.
Mendeleev’s periodic table helped scientists make sense of chemical behaviors and helped kick off a new era in physical chemistry.
2X2 decision making matrices used by management consulting firms (e.g. Ansoff’s Matrix, the BCG Matrix) are ways of sorting situations into types and figuring out what the best strategic decisions should be for your business. A system containing a range of “types” is called a “typology”. Customer segmentation and profiling can produce typologies that can be used in sense-making and strategic decision making.
Your challenge is to identify the most salient perspective for the problem, at hand, capture those perspectives into frameworks, and then help their users seek insight from interacting with the frameworks.
You have areas of your organization where new ideas, opportunities and risks for your business come from.
Traditional taxonomies do not work effectively in these areas, simply, because these areas are shifting, poorly understood, and changing all the time. A Discovery taxonomy project uses taxonomies that are extremely flexibly to seek usable insights from within these areas.
For example, R&D teams often brainstorm maps of science and technology using their own implicit science and technology taxonomy structures, but using their own implicit science and technology taxonomy structures, but then disrupting these taxonomies by making connections across disciplines, or combining concepts at random in unexpected ways. For example, how can an algorithm that predicts cracking in the earth’s crust as a result of an earthquake, be deployed to develop an anti-cracking cosmetic for skin?
Your role in a Discovery project is to be able to supply multiple taxonomies and ways of visualizing combinations of concepts, e.g. through system maps.
With an understanding of these five taxonomy strategies, it becomes clear that as you move up through the different project strategies, the taxonomy itself starts to take second place to the processes that you need to build and follow.
Taxonomy work usually involves more than simply creating a taxonomy. It can often also be about brokering common language, helping different communities communicate with each other without giving up their identity, and supporting your discovery processes.
So, which is the best taxonomy strategy for your purposes?
Tell us about your efforts to create taxonomies in your organization.
What lessons did you learn in creating your taxonomy?
I will be speaking at the following events:
June 5th– 8th, 2012 AIIM ERM Master class in San Francisco, CA
June 12th, 2012 Info360 ECM Practitioner Pre-Con in New York, NY
June 19th – 22nd, 2012 AIIM ECM Master class in Houston, TX
#BigData #ElectronicRecordsManagement #ERM #TaxonomyandMetadata #ECM