Confusing nodes and navigation
At this stage it is important not to confuse a taxonomy with a navigation scheme. The labels used on a navigation bar may be different to the underlying taxonomy nodes, and this difference should be made clear to all the people you will be working with to complete the task.
It’s a common misunderstanding but, in this article, we will cover taxonomy in terms of site structure and organisation and taxonomy in relation to navigation and design. The two aspects are unique and their relationship with a taxonomy is very different. One is organisation of content and the other is the presentation of its organisation.
Principles of Information Architecture (IA)
When creating a taxonomy, there are several aspects to the task that need special attention. A good place to start is looking at the constituents of the practice of information architecture (IA). When a solution is devised using IA methods, three factors determine the outcome. These are the users, the business (context of use) and the content that is being ordered.
Before using these areas as a basis for taxonomy creation, some awareness of the business culture needs to be applied as well as a realisation that, although one person may be responsible for the creation of the taxonomy, it is necessary to have buy-in and collaboration from a large group of colleagues and users.
Several tools allow this to happen. The diagram below details the elements that combine to form a cohesive, holistic view of a taxonomy that needs to be created.
Fig.1 – Ingredients of a Taxonomy
Looking at the diagram it’s easy to see that taxonomy (for want of a better word) is actually a combination of complex tasks that have different timelines and use different teams to collate the information.
The job of a taxonomist is to bring all these elements together in one place, to define how the taxonomy should be presented, and to ensure that the underlying information is there for development teams and the business to use going forward.
Understanding the content
To gain an understanding of our content we need to look in detail at the following items;
- Content Audits – This is a technique of looking at every major section of a site and inputting it into a spreadsheet. Although a laborious process, it gives you very detailed insight into how the information on the site is organised.

Fig.2 – Content Audit
- Industry Taxonomies – There are several taxonomies that are available for public use and do form a good basis for a quick way to produce a straw taxonomy (a draft). If there are some available, use them to help classify your content as the terms used here will probably have been reviewed and tested.
- Search analytics – A great way to see how your users think of your content is to check on the terms that they search on. This is a very revealing way to understand the terms used to access areas on your site.
- Social bookmark sites – some sites, such as del.icio.us, show clouds of user generated tags around industry labels. This is a good way of seeing if what the users’ say they want tallies with what they actually use in terms of categorising the site’s content. User generated tags have a role to play in formulating the holistic taxonomy strategy (that incorporates the user, business and content)
- Content card sorts – by showing the content types and the category labels, this type of closed card sort can help in defining what a user expects to see behind certain labels. Open card sorts are another approach that allows the user to also decide the categories to rationalise a website’s content.
- Content mapping or clustering – this technique involves pooling groups of content into areas and seeing what emerges. They are likely to be along a theme, or ideally a node in the taxonomy. This type of content mapping is an important element of building a new taxonomy. This is where industry experts can help in giving a feel for content and how this should be reflected in the taxonomy and eventually in the visual design.
- Web analytics overlays – Most web metric programs have the facility to see the clicks that users are making whilst interacting with the site. By doing this, we have an idea of how they perceive navigation elements and their labels. This is vitally important when we come to designing a navigation scheme that reflects the taxonomy.
Understanding the users
Outside the immediate business context, you need to rely on user studies. Personas can be a great snapshot of how different user groups think of the content they consume and how they interact with it. The best form of research is ethnographic studies ,but even an online card sort with a select user group is a great way to harness the collective view of the user.
By placing the persona at the heart of the taxonomy formulation, you are ensuring a level of user centred focus that becomes the fabric of the website – its structure. Looking at qualitative and quantitive data will yield a solid grounding for the taxonomy.

Fig.3 – Analytics form a good basis to understand actual user behaviour
To gain an understanding of our users we need to look in detail at the following items:
- User behaviour – from looking at web analytics we can see patterns of behaviour that emerge from the user’s interactions with the site. An even better way is to study users for a fixed time period and see how they interact with the website and how they navigate to areas within the taxonomy.
- Depth Interviews, online surveys – by asking questions around the taxonomy and how the site classifies its content, we can gain a true understanding of how users think content should be organised.
- Task analysis and customer activity cycles – also give us clues about the user’s working life and how the website fits within their mental model of how our content should be organised. This is from more of a task flow analysis view than a personal view, but is very useful to understand the business context in which they are working.
- Personas – a highly memorable collation of the user’s specific needs around a fictional character that is based on factual research. Personas can symbolise how different user groups perceive the content that we are trying to organise. In short, all the findings of the elements above help build this definition.
Gaining understanding from the business context
Unlike the users and the content, the business context is a really hard area to define with rigid tools and techniques. The problem is that each context is likely to be affected by a pervading business culture that will have a direct effect on the quality of work that will be produced.
To gain buy-in from key people requires a charm offensive and a very good elevator pitch that encapsulates the vision of the site in simple, easy to understand language. It is this vision statement that will underpin your discussions about the site and its taxonomy. Hopefully, before you embark upon this exercise, there will already be some work done, detailed below…
- Proposition Plan – a detailed statement that has been circulated and approved helps no end in defining the scope of the project and how far the taxonomy needs to reach. This gives you some good parameters to work to.
- Website Strategy – this is an overarching document that talks of how the site will develop over time. The importance to the taxonomy here is that it will give you an idea of how the taxonomy will need to adapt to change as the strategy comes into reality as time passes. To produce an adaptable taxonomy is a goal that should be achieved during the design process.
- Competitor Analysis – One of the first things to gauge how the taxonomy should look is to see how the competitors organise their content. If it is good or bad, pick up on mistakes and successes.
- Content Road Map – An extension of the website strategy, this is the actual editorial content that will be produced to align with the vision of where the site needs to be directed in the future.

Fig.4 – Concept model diagram
- Concept Model – This diagram encapsulates the elements above in an easy-to-understand visual explanation. The concept model is a snapshot of the strategy at this stage and also makes clear the areas which a taxonomy will have to address.
It is important to note that where taxonomy design struggles is often where stakeholders do not fully understand, or agree with, the strategy of the site moving forward. The best taxonomy design can often be ruined by a lack of engagement from key people on the team. Without the input of all available knowledge, it is likely that something will be missed and the structure will be faulty.
Starting Taxonomy Creation
Taking all the components of taxonomy creation into account, the way it must be brought together is a very detailed iterative process. The mapping process begins by looking at content and then aligning user needs. It is probable that you will have a massive amount of data to deal with as you work your way through the elements of a taxonomy.
Remember here just to categorise content and not be concerned with the types of content you have. The navigation scheme can be dealt with later. But try to align users, business and content simultaneously as you work through the content and use as many of the techniques and documents listed in the first part of this article as is possible.
The Straw Taxonomy
It is likely that the creation of the taxonomy will come down to one person, although many different parties will want input into its creation. The team can meet informally but regularly.
The naming of elements is fraught with politics: a way to get around this is to rely heavily on user research. Card sorts, open or closed, can give you a good feel of what your users, editorial or sales teams would like to call different labels. Gather the people who are experienced in the domain and simply work through a draft taxonomy that you have formulated by looking at the content types of the website and interviewing journalists and business owners. Just discussing their view of the industry is valuable grounding to producing the straw taxonomy.
When you present the straw man, expect to get it burnt. The point of these workshops is to thrash out misconceptions with those who know the domain better than you or who can bring another perspective to the task.
It is also in these sessions that you will become aware of the corporate culture in which you need to operate and this may be difficult. However, if you feel the meeting is being railroaded, always pull the meeting back to the user and qualify it with research to back your views up.
Iterate and Test

Fig 5. Teragram classification software where rules written by a taxonomist classifies the content
Once amendments have been made, revisit the taxonomy with the same team members and also bring in other parties. For instance, sales employees will bring a different perspective than editorial employees. There may be financial considerations that need to be built into the taxonomy to ensure that the site is a commercial success.
Until you reach agreement with all parties and the taxonomy is defined, you cannot start on the navigation scheme and decide how you will display the terms to the user.
The Controlled Vocabulary (CV)
During the process of creation you will discover associative words that belong to the main term. Always make a note of these, as they form the controlled vocabulary that needs to accompany the taxonomy. The different types of words are listed below along with their usage
The CV terms are also really helpful for search engine optimisation (SEO) as they are the keywords that users enter when talking about the category type. The CV details ‘aboutness’ around the taxonomy nodes, it gives descriptions, preferred and non-preferred terms.
This forms the basis to test the taxonomy against your content using a categorisation engine – a computer program that classifies content by using linguistic rules to place content in the relevant areas. The results of these tests will determine how good your taxonomy is at reflecting the content that it endeavours to describe. Don’t worry if you do not have the skills to run these tests as it is something your web development team should be able to undertake with help from the categorisation software company.
Fig 6. Relationships between different terms in a controlled vocabulary
Each taxonomy category is named using a formal label. This is what your project team will have decided to call this category / taxonomy node, also known as the Preferred Term.
- Information for User is available to guide browsing (and searching).
- Scope Note: to guide the designers of the rules which will control automated categorization of content to this category. When using a categorisation engine, such as Zibb, these terms help it to classify content to the right parent node within the taxonomy.
Non-Preferred Terms are synonyms and alternative names. Included in the Non-Preferred Terms are candidates for labelling. The Non-preferred terms will also ultimately include common misspellings, plurals etc. to enable users to land in the right part of the taxonomy structure and to assist SEO.
Related Terms: indicates other nodes in the taxonomy to which there is a notional link, if not a formal one. This relationship is important to allow serendipitous browsing. For instance digital cameras could have the relational term of Camera Phone.
Search Terms: Words and phrases pertinent to this subject which can be used when creating rules for automatic categorisation. These are not listed in any order of importance nor have they been tested, but should be considered as candidates for rule creation.
These search terms are not necessarily the same search terms (from seed keywords or social network site data) which may give rise to the creation of the taxonomy node.

Fig 7. An example of the relationships within the Catering industry
Final Schema
Once the taxonomy and CV have been finished and been tested on new content (to ensure the taxonomy holds up to changing content) you can be confident that the taxonomy is robust enough to be rolled out.
Keep in mind that a taxonomy will evolve as the content it describes changes over time. Refer to this article for advice on maintaining a taxonomy.
The final stage is to now apply the taxonomy nodes to the navigation schema. Ensure that the different labels reflect the underlying nodes, but do so in language the user understands. The results of user research will help here.
Fig 8. The step by step guide of creating a user centred taxonomy
Creating and Maintaining Taxonomies: Step-by-Step Summary
1. Assemble the team – include business owners, sales, content providers, user experience, marketing and interaction designers
2. Scope out the taxonomy requirements – elements gained in understanding the business will help here
3. Analyse data – take as many sources as possible from understanding the users and the content
4. Propose a taxonomy – take your straw taxonomy to the project team
5. Iterate and test – make changes and run tests against the content, both old and new
6. Compile a Controlled Vocabulary – invaluable to SEO and the rules writers for automatic categorisation software
7. Launch and maintain – Monitor the ability of your taxonomy to adapt but expect new nodes to be brought in and redundant ones consolidated over time.
Summary
When building a robust taxonomy, there are many different elements and tools of which you should be aware. You also need to think about how the taxonomy will be displayed to the users and how it will fit within the website design.
Developing a site taxonomy with an interaction designer is a very helpful way to do the design, because the way the site is structured directly affects navigation and interface elements. Interaction designers can be a great sounding board and give immediate visual feedback.
The taxonomy of a site and its structure are the most important areas of website design and a site that fails often has a poor user experience around this area.
Get this right and you will be on course to ensuring that your users will have a quality experience because their needs are within the fabric of its design.
Tags: Taxonomy, taxonomy design, Taxonomy Management, Teragram, User Centred Design







Hi, James!
This is an excellent work about taxonomy, please let me call it ‘taxonomy 101′ in new period
We would like to translate this article into Chinese version on IAChina Website for sharing your good ideas. Ask for your permission. If possible, please email us your answer. (I’am sorry to leave message here, because I can’t find your email address)
Regards,
Van