Introduction

In the time that I’ve been Programme Chair of Taxonomy Boot Camp London, I’ve noticed a healthy increase of interest in the business use of taxonomies, ontologies and knowledge graphs. Leading the development of the programme means I encounter many examples of good practice and excellent real-world implementations. Taxonomy is well and truly breaking out of its roots in the library and information science worlds to find uses in all sorts of digital applications.

In the private sector, taxonomies are business assets, constructed to support products and services that are intended to stand out in their marketplace and, one assumes, earn revenue for the organisation. Taxonomists want information management to be recognised as a core business process, and as something of great value. If people can’t find what they’re looking for on your website or app, they will go somewhere else. Very few companies have totally unique and proprietary information or data to sell.

In cases where taxonomies are used to help users inside an organisation find and use information, these are often developed because an organisation wants to reduce inefficiency. Taxonomies in the public sector may be needed for either internal or external use, but again are built in order to achieve a specific outcome that benefits the business.

Here I explore some of the reasons why taxonomy projects fail and what taxonomists and others can do to ensure they don’t. The issues and mitigations discussed are equally applicable to enterprise search and ontology/knowledge graph projects. The approaches described are designed to be useful whether someone manages a taxonomy as part of a wider role, is a dedicated in-house resource, or is a consultant. I also share insights into award-winning taxonomy projects.

The stages of a taxonomy project

There are many potential bumps in the road between a taxonomy project being conceived, and the taxonomy being a long-term success. It’s rarely a smooth or quick process to get a taxonomy or search project approved in the first place. Once the project is completed, the work of embedding a taxonomy properly is usually part of a wider change management initiative. Metaphorically flinging a new taxonomy ‘over the garden fence’ and expecting users to adopt it from day one, without helping them to understand the business benefits, rarely ends well. In the end, without an ongoing plan to keep the taxonomy for its intended purpose (or purposes), then the investment of time and money will have been squandered.

Business approval

Starting at the beginning, there are many ways that taxonomists can bring colleagues along with them on the ‘journey’.  It comes down to listening, educating and advocating as much as possible within the organisation. Every conversation is an opportunity to sell the benefits of taxonomies, not just at the senior management level from where a project sponsor might emerge, but also at the peer level, where colleagues might be wary of how a taxonomy could change their current ways of working.

(One thing that can happen during the project is that the sponsor changes, causing uncertainty and even risking the completion of the project. Business priorities may change as a result. Sponsor changes are outside a taxonomist’s control. Other than continuing to cultivate understanding among a range of senior managers, there’s not much we can do with this one.)

Finishing the initial phase of development

Implementing a taxonomy, whether on its own or as part of a larger technology project, is a reason for celebration. The hard work of gathering requirements, understanding user needs, understanding the domain and analysing the content or data has been done. This knowledge has been translated into a live, working taxonomy. Everything is cool now, right? Not necessarily.

The most common failure scenario is that once the taxonomy project team has completed its mission, it is disbanded without a commitment to ongoing maintenance of the taxonomy. The project checklist item ‘do the taxonomy’ is ticked off and not enough consideration is given to day-to-day operation and the maintenance of quality.

Moving into business-as-usual

There is a common misconception that a taxonomy is just ‘finished’, yet most domains can and do change over time. Terminology moves on, new entities and ideas emerge. Examples of this include how the acceptable language for certain mental health disorders has changed over time with greater social understanding. Even fairly settled vocabularies such as ‘the capital cities of the world’ have to be updated when countries rename or move their capitals (for those who care, there have been two such recent examples, namely Kazakhstan and Burundi).

Quite apart from simple factual inaccuracy, there may be wider societal issues of bias and terminology choice to consider in many domains. Some organisations are building vocabularies that allow for more than one preferred label, because there is no one accepted name for say, a mountain or piece of colonised land.

For these kinds of reasons, it is always advisable to have in place a role (or a part of someone’s role) to oversee the taxonomy. Yet even this may cause a problem after a taxonomy is deployed. This is because the taxonomist (assuming there isn’t a wider team) becomes a Single Point of Failure. When they change companies, retire, or just move job inside the organisation, their knowledge and enthusiasm may not be replaced. (Of course, many companies do keep their taxonomist role filled, but as an external taxonomist I am more likely to see those places where they haven’t done this properly!)

Projects can fail even if those in charge fully understand the value of taxonomies in supporting broader business objectives. Managers can have an ambitious vision for using knowledge graphs to power data-driven products and services, yet still not care enough about the quality of the underlying vocabularies and structures. An example of good practice was Electronic Arts (EA)’s presentation at Taxonomy Boot Camp in 2019. The team emphatically cares that the right taxonomies and ontologies get built. The team takes the long view so that future enterprise-wide requirements are borne in mind as well as immediate project needs. Taxonomists form a core part of a team that also works on ontologies, knowledge graphs and content models.

Models, such as the ones EA are building, depend on agreed and shared definitions of the types of entities or things the organisation cares about. For example, in a recent role I was trying to model agreed definitions for digital asset types (e.g. video trailers, background images, call-to-action text). Everyone I talked to agreed there was a need to standardise and reduce the amount of types people were creating. This was because there was rampant inefficiency, duplication of effort, poor communication between production teams and poor findability. Without buy-in from the people overseeing these teams, the processes and systems won’t change, and new vocabularies and models won’t be adopted. Eventually this may either lead to remediation work being commissioned, or in the worst-case scenario, a total loss of business confidence in using taxonomies at all.

Longer-term sustainable maintenance

Helping others to understand that business-as-usual is as important as project work is a vital task. The taxonomy must be kept relevant and useful for its end-users, and this only has a chance of happening if care and attention are given to ongoing maintenance.

Governance is an important part of this. This can mean anything from a designed framework that encompasses committees and review schedules, to a pragmatic commitment that one person will be responsible for looking after the taxonomy and communicating with others about proposed changes.

It can be easy to fall into the trap of focusing on the ‘shiny new thing’ that comes out of a project and to underestimate the value of low-key day-to-day deliverables. These include search log analyses, data analytics insights, tagging audits, reporting on taxonomy change requests, and governance processes. All of these things give precious insight into what’s working and what’s not and may even throw up new and surprising findings. Maybe a whole section of the website is rarely visited, or users are searching en masse for a term variant that no-one thought of. Maybe a taxonomy term is getting disproportionately tagged against content, and it’s only because it’s the first entry in the autocomplete list for a common word.

Working across disciplines and business siloes for long-term success

There are strategic and practical things that taxonomists can do to help the long-term success of projects. Getting and retaining business buy-in is arguably the most important. There might be one person who is project sponsor, and this is critical, but it is also important to try to build relationships with others across the organisation. After all, taxonomy (and search too!) is not something that fits neatly into one business function. It crosses disciplines such as technology, content strategy, design and user experience, product management and change management. Senior stakeholders in all of these areas should be supported to understand the value of taxonomies.

It’s clear that input does not just come from the taxonomist. Other disciplines can and should be involved, e.g. content designers who understand structured content and markup, or developers who understand tagging beyond a simple view of ‘stuff added to a piece of content’. I would like to see organisations treat taxonomy, metadata, search and tagging skills are as core to their overall set of digital skills.

By way of analogy, it’s increasingly common to observe that professionals who aren’t digital or content specialists are expected to contribute content about their particular area to digital workplaces. I would like to see organisations encourage staff to interact more with, and understand more deeply the value of, taxonomies. The best colleagues I’ve worked with over the years had the curiosity and imagination to deeply understand how taxonomies fitted in with their own work and thus benefitted the wider project.

It’s great that things are moving in the right direction (after years of ‘is taxonomy obsolete?’ blog posts!) But it will be even better when people at all levels of an organisation, and across all sorts of digital disciplines, are fully on board with the excellent work that taxonomists are already doing. Every year we see this superb breadth and depth at Boot Camp, it’s time the wider digital industry understood this too.


This chapter was originally published in Search Insights 2021. Download the report for free here:

Search Insights 2021 by The Search Network
Search Insights 2021 – ENGLISH
Search Insights 2021 – FRENCH TRANSLATION

Leave a Reply