This is the fifth article in the series ‘The Art and Science of Taxonomy Development for Market and Competitive Intelligence’. Click here to view other articles.
A thoughtful taxonomy helps users easily navigate to the information that is relevant to them. It should allow for accurate and consistent tagging of information with terms or tags that are easy to understand. But the English language is full of exceptions and ambiguity, and therefore a fertile breeding ground for confusion. And when users get confused, they stop using the information system. One of the success factors for any information system is how well its users can comprehend the underlying meaning of a term in taxonomy. Based on our experience, we recommend the following guidelines for naming terms (tags) that will help in developing a long-lasting and maintainable taxonomy.
1. Tags should be clear and unambiguous
A taxonomy tag ‘French’ could imply the French language, French wine or the people of France. A ‘bank’ for depositing money or the ‘bank’ of a river? Ambiguity occurs when a word or phrase has more than one meaning. The goal of a good taxonomy is to eliminate the guesswork for its users. This requires that each taxonomy tag corresponds to a unique and distinct concept. However, if there are no alternatives to using such ambiguous terms, we should explicitly clarify the term using some kind of qualifier. Following are a few recommended methods:
- A. Qualifiers
Use a qualifier such as a parenthesis, to add the context and disambiguate the term. For example, French (Language), French (Person), or French (Wine).
Caution: Use of qualifiers should be the second preference (not the first) because of the problems that they can cause in indexing (storing) and searching (retrieval) of information. Search engines could give inconsistent results while processing qualifiers like parenthesis, hyphens, ampersands, or punctuation marks.
- B. Compound Terms
Rather than using a qualifier, whenever possible, combine the two terms to form a compound term. For example, use ‘Strategic Initiatives’ rather than ‘Strategy (Initiatives)’. In Contify (when no other option is available) we use such parentheses to display the context along with the tags.
Note: Qualifiers should be standardized and consistently used across the taxonomy. This should be documented in the taxonomy governance document.
2. Tags should be mutually exclusive
Each tag in the taxonomy should represent a single concept (or unit of thought). Therefore, two tags should NOT have an overlap in their scope definitions. For example, it is a bad idea to include tags like ‘Customer Wins’ and ‘Business Expansion’ in the same taxonomy. Does ‘Business Expansion’ include ‘Customer Win’ or not? In addition, does ‘Business Expansion’ include ‘New Products’? Office Openings? Partnerships and Alliances? Tags with overlapping meanings cause confusion. Because of this confusion, users may tag similar information inconsistently, and the end-users could get confused about which tag to select while looking up a competitor’s sales wins. If users get incomplete information or don’t get the information that they expect, then they will stop trusting the entire information system. It’s not their mistake — it’s the result of a poorly designed taxonomy.
Caution: In order to get mutually exclusive tags, we should not swing the taxonomy to the other extreme and start adding highly specific tags for every piece of information. It will introduce another problem – discussed in the next point.
3. Tags should include Most but not All possibilities
For example, in addition, to ‘Positive’ and ‘Negative’ sentiments, should we also include ‘Neutral’? Location taxonomy should include Countries, or should it also include States? What about Cities? It is tempting to develop an exhaustive taxonomy with a unique tag for every piece of content. Such highly specific tags for the sake of a complete taxonomy leads to a bloated taxonomy. This makes the taxonomy difficult to manage and difficult to use by the users. We shouldn’t strive to achieve perfection and absolute granularity. The goal is to get a good-enough taxonomy to start the process. Give the taxonomy room to grow with the information, users, and use cases. Based on our experience, we recommend the following guidelines to control the taxonomy terms around the core use-case:
- Map the terms with the use-cases. Avoid adding a term if there is no direct use case for it.
- Taxonomy should be extensible to accommodate new information. Design it for iterative development.
- Develop a taxonomy governance document to manage how changes will be made, tested, and communicated to the users.
Don’t try to develop a be-all and end-all taxonomy in the first attempt. For example, Contify’s initial standard taxonomy had a tag for “CEO’s Compensation”. There were a few news updates about the compensations and we couldn’t tag those with any tag. Therefore, a new tag was created — a bad idea. To properly use a taxonomy, users will have to memorize it. If they can’t remember terms, they won’t use them. The result is inconsistent tagging of the content, and users unable to retrieve the information. Consequently, they will lose trust in the information system and the whole organization will be back to square one — an information mess. We have to make it easy for our users to use the taxonomy.
4. Tag names should be intuitive and reflect the scope
‘Partnerships and Alliances’, ‘Financial Results’, ‘New Offerings’ are good examples of tags because their names themselves convey the information associated with the tag. But tags like ‘Strategy’, ‘Digital’, ‘Competitive’ are not good examples because the name does not unambiguously communicate what’s included in these tags. If there are no better alternatives, then make sure that a scope note (discussed in the next article) describes the inclusions and exclusions. This scope note should be easily accessible to the users of the taxonomy.
Caution: Users, usually, will go by with the face value (name) of the tag. It’ll be a mistake to assume that users will read all the scope notes and will memorize what’s included and excluded from the tags.
5. Tag names should conform to a style guide
Be consistent with the names of the tags. For example, use the same grammatical form, plural versus singular, company name with the legal entity or without, abbreviations or spell out the complete term, multi-word terms, capitalization (upper, title, capital or sentence), American or British spellings.
There are no absolute right or wrong rules. The key is to have a consistent schema for naming the tags. It makes it easier for users to understand the taxonomy, and by extension, the architecture of the entire information system. It also allows multiple owners to further develop and manage the taxonomy.
6. Use Nouns (not Verbs)
A taxonomy term should be a noun — a word used to identify a single or a class of people, places, things, or abstract ideas. The condition is that they have to follow a consistent and predictable structure. The primary purpose of a taxonomy is to achieve consistency in the description of information so that it can be stored, retrieved, and shared easily.
Avoid the use of catch-all wildcard tags like ‘miscellaneous’ or ‘others’ — it weakens the whole taxonomy. It wrongly encourages users to start using these ‘shortcut’ tags rather than treading the often longer route of either understanding the whole taxonomy or updating it with the new tags. It is better to leave some information without any tag than bloating the taxonomy with hyper-specific tags or adding ambiguous tags like ‘Miscellaneous’ or ‘Others’. We should trust that even if there is no specific tag, users can find the information using a keyword search.
Best Practice: Write Scope Notes
Write the scope note as you are developing the taxonomy. Just the act of writing the scope note forces us to think through the different aspects of the term, which avoids confusion and saves a lot of time later when other users start using the taxonomy. A scope note should clearly state the chosen meaning of a term — what’s included and what’s excluded. A scope note ought to be viewable to all. Keep it crisp and concise to facilitate a quick reference by the user. Do not make it superfluous by describing aspects that are already self-implied from the tag’s label.
Indicate other meanings of the term that are commonly used in natural language, but that is deliberately excluded from the taxonomy. In case of possible confusion, explicitly write what’s NOT included in the tag. For example, the scope note for ‘cybersecurity’ for a semiconductor manufacturing company could read: “Includes cybersecurity within manufactured products such as cars, phones, semiconductors, manufacturing plants, etc. It does not include cybersecurity of banks/financial institutions.” In Contify, we display the scope note when the user mouses over the tag.
These guidelines should be explicitly spelled out in the taxonomy governance document. For example, IKEA has a meticulous internal logic to naming scheme for its over 12,000 products. It also offers a peek into Scandinavian culture. Desks and chairs are given Scandinavian boys’ names; curtains and fabrics are given girls’ names, outdoor furniture is named after Scandinavian islands, rugs after towns in Denmark and Sweden — so on and so forth.
After the candidate terms are selected and named properly, we can now start the process of building the taxonomy.