Taxonomy Design Best Practices
Over my years of taxonomy consulting I’ve often been asked to codify the best practices regarding taxonomy design. A simple search online shows that this is a common question, though one where a lot of the current answers appear outdated. Taxonomy design, as well as the information systems that leverage taxonomies and the business needs that drive them have evolved over the last several years.
Taxonomy management tools and auto-categorisation systems are more mature and easier to use. Many content and document management systems (and really, the broader set of information management products as well) have grown into suites of products through a multitude of acquisitions by industry leaders, while open source upstarts have introduced some much needed volatility and new functionality into the market. Most importantly, businesses themselves are much savvier regarding the value of their information, the needs to use and reuse it, and the reality that everyone is not just an information consumer, but a potential information creator as well.
As a result of these dynamics, many of the taxonomy best practices from a decade ago have shifted or changed. Others are as critical as ever. The below represents the core set of today’s business taxonomy design best practices.
Define and Document Your Purpose – Taxonomies, despite a lot of improved understanding over the last decade, can still be somewhat esoteric to the non-indoctrinated. As such, I too often see a stakeholder group that is confused about the need for taxonomy and the value the taxonomy will provide. Every taxonomy design effort should begin with a clearly documented and shared understanding of the Who, What, and Why of taxonomy. Who is our audience? What are we “taxonomising” or tagging? Why are we doing it (What’s the business value that will be derived?)?
Focus on the Business User – I often tell my clients that this is the only best practice that is non-negotiable. Every design decision should come down to that which would best serve the end user. This is also why the development of personas is so important. An early stage in any taxonomy design effort should be the identification of your audience/users and a clear understanding of your “lowest common denominator” user. If you can design for that person, you will best serve the interests and needs of all your end users, be they customers, the public, your own employees, or a mix.
Understand your Publishing Process – Depending on the type of system, type of organisation, and, purpose for the system, the content creation and publishing workflows can vary wildly. Some systems have content published only by a few, full-time content strategists. Others have a fully democratised content management approach where any user of the system also has the ability to publish into it. The number of tags you can expect to have consistently and correctly applied to your content is heavily dependent on who will be applying those tags. Systems with a broad base of content publishers must sacrifice some level of taxonomy and tagging granularity for overall usability. I will always take a smaller and simpler set of tags that are consistently applied over deep and detailed taxonomy that is used sporadically. The use of auto-categorisation tools also comes into play here. A system leveraging auto-categorisation (if designed and implemented properly) will decrease tagging time and increase consistency, allowing for a more granular taxonomy design.
Use the Simplest Language Possible – One of the best taxonomy consulting services I can provide an organisation is sometimes to serve as the outsider that doesn’t recognise their internal jargon and acronyms. Internal taxonomy designers too often get overly comfortable with their own terms and concepts. I always encourage organisations to “burn down” their language into the most simple and cleanest terms possible. It is this language that will most make sense to a first-time visitor to a site, an employee on their first day, a beginner, or any other lowest common denominator. As an added value, this clean language is also the easiest to translate for internationalised taxonomies and the easiest to maintain in order to minimise long-term costs and administrative burdens.
Deconstruct your Taxonomy – One of the keys to today’s taxonomy design efforts, and one of the major changes from the past, is that the concepts of faceting, along with advances in taxonomy management and information management technologies have given us the ability to step away from the “one taxonomy to rule them all” model. As opposed to a terribly deep and detailed taxonomy with a mix of different terms, today we strive for clean and simple taxonomies, each of which may power a separate metadata field in a one taxonomy for one metadata field type of model. This easily enables the concepts of faceted navigation and search now available out-of-the-box in most information management systems. It also improves the accuracy of auto-categorisation tools and creates the potential for much easier taxonomy and content governance by ensuring mandatory fields are completed correctly. Moreover, a deconstructed taxonomy design yields simpler and flatter taxonomies that are much easier for the average business user to leverage and understand.
Leverage the Wealth of Information that Already Exists – Thanks to today’s analytics tools, most organisations have a pretty good understanding of user behaviours. Understanding what words people are searching on, what information they’re accessing the most, and how they’re navigating a site are all extremely valuable tools to taxonomy design. For instance, recognising a particular term has been searched upon more than any other can be an important key to recognising that term should be included in a core taxonomy design. Equally, knowing what your most sought after content is plays a critical role. I often work with organisations to understand that and then focus on a first implementable version taxonomy design that primarily serves that “MVP” content.
Plan for the Long-Term – We all know that no taxonomy is ever finished. An organisation’s needs and strategies change, as do their mission, services, products, and employees. As a result, content is constantly in flux and a taxonomy design must be adaptive in order to address these changes. Moreover, on an average taxonomy design effort, the world of potential users will not be able to respond to it until after it is deployed. Every organisation needs to have the resources in place to capture and respond to both the active and passive feedback that will come after rollout.
Leverage Governance – Though taxonomy governance many not be the most exciting topic, it is by far one of the most critical to long-term success. Taxonomy governance will ensure a design evolves to better reflect the needs of the business and users, but does so in a sustainable manner that doesn’t “break” the original design and functionality.
Look to Usability Best Practices – The concepts of taxonomy design and usability are rather similar. When designing taxonomy, I always encourage my clients to consider usability as well. As more and more taxonomies are “front-facing,” leveraged directly for navigation and findability, this becomes even more important. Even “back-end” taxonomies need to be usable for the taggers. Traditional views of navigation and “clicks to content” hold true for taxonomy design. The days of six-level deep business taxonomies are long gone, or at least should be.
The above are the taxonomy design best practices that matter most in today’s world. The core theme that runs through them is the focus on the business and the user, hence my use of the term business taxonomy. Focus on practical business value for the business and business users, and your taxonomy design effort will be off to the right start.
Zach Wahl is Principal, Knowledge Management and Taxonomy at Washington, US based consulting firm Enterprise Knowledge. Email him at zwahl@enterprise-knowledge.com