Carlos García-Egocheaga, CEO of Lexsoft, explores the important role of taxonomy in KM project success.
Firms tend to use the terms taxonomy, classification, glossary and keywords interchangeably when building their knowledge management (KM) systems. It’s understandable – the differences between these terms are nuanced. However, a lack of appreciation of their correct connotation – especially of taxonomy – could hamper the design of the KM system and therefore successful adoption of this business function in the firm.
Taxonomy is not classification, glossary or keywords
Foremost, the most common confusion is between the terms classification and taxonomy. It’s true that the terms classification and glossary are components of taxonomy, but the latter is much more. For a start, classification and glossary are static in nature, whilst taxonomy is dynamic. Taxonomy is a layered structure and also works as the filter within the workflow that lawyers use to navigate through content. Let’s illustrate the concept using Amazon search. If you search for “mobile phone”, Amazon will throw up over a thousand results, alongside filters to help refine the search – i.e., by price, brand, customer reviews, camera resolution, internal memory, and so on – and enable the individual to eventually find the exact product he or she is looking for. In the KM system, the taxonomy serves as the filter in this Amazon search scenario.
Likewise, often people substitute “keywords” with taxonomy and vice versa. This leads to duplication in the KM system because the terms used as keywords can and often are also the terms for which taxonomies are built. To illustrate, VAT, real estate, Brexit and Spain could be keywords, but taxonomies could also be developed to address each of these areas.
Therefore, it’s important to recognise the role of keywords and taxonomies within a KM system. Keywords are free-text terms that are best kept as very specific classifications for which a separate taxonomy isn’t required. Take the terms “Contract”, “Madrid” and “M&A”. These can be either keywords or part of a taxonomy, but a particular firm could classify “Contracts” as part of the “Document Type” taxonomy and “M&A” as being part of the “Commercial Law” taxonomy – leaving “Madrid” as a free-text inputted keyword.
At a future date though, if the firm’s activity in the Madrid region expands, “Madrid” could be removed as a keyword as it becomes part of the “Jurisdiction” taxonomy. For this reason, if a firm has a long list of keywords, an approach might be to replace the terms with a set of predefined taxonomies so that the lawyers can use them as filters while searching.
Key considerations for taxonomy design
With taxonomy playing a “make or break” role and due to its complexity, there are some key considerations that are worth paying attention to when designing taxonomies in the KM system:
- Focus on the firm’s core business. If the majority of your firm’s business comes from providing tax advice, it makes sense to concentrate on the tax taxonomy in your KM system, in the first instance – subsequently branching out into related and other taxonomies for the various practice areas. For example, as a Corporate Law specialist, if the majority of your revenue comes from M&A activity, creating a taxonomy for M&A first and then moving to Corporate Governance, as the next biggest revenue generator would be a logical approach.
- Ensure a familiar look and feel. In addition to referring to internal resources of information and knowledge, lawyers routinely access external sources such as Wolters Kluwer, vLex, LexisNexis and such. When designing your taxonomies, deliver a similar “look and feel” to these sources. Familiarity will encourage the adoption of the KM system. It’s the same reason some of the best legal technology vendors develop their solutions in a manner that offers lawyers the familiarity of Microsoft Office tools.
- Restrain the urge to create “depth for depth’s sake”. Frequently we hear discussions at firms on how deeply their taxonomy should be designed – five levels, 10 levels, more?. This is an immaterial conversation as taxonomy design must be dynamic so that the KM system delivers Amazon-like search capability, as described above. Instead focus your attention on ensuring that for every level and term in the taxonomy, the KM system houses the right documents and content.
- Create a robust “search” strategy. Before building any taxonomy, define your search strategy. If you are after Amazon-like search capability, you will need to build a deep taxonomy that will provide a large variety of dynamic filters. On the other hand, if you opt for a Tree Navigation type search approach, a simpler taxonomy works a lot better. A well-defined search strategy will ensure that every time lawyers do searches in the KM system, they always find content underneath that search.
- Build-in reporting. A KM system is only as good as the content residing in it. You can have the best designed and structured taxonomy in concept, but without the corresponding documents, it’s of no use. Embed reporting as a core part of your KM system design. You will have insight into things like how many documents reside within every single taxonomy? Who are the top contributors of documents for the KM system? Which documents are accessed the most? How long does it take from submission to publication of documents? And so on.
More crucially though, the insight will help to continuously fine-tune the KM system. You might find that there’s minimal usage of X taxonomy, suggesting it’s no longer required. Conversely, lawyers may be routinely searching within that taxonomy, but there are no new documents being submitted for inclusion – giving you an opportunity to proactively discuss and address the problem with Practice Heads.
These considerations will ensure that the taxonomy design meets lawyers’ needs, and is in tune with their way of working. Today firms are looking to adopt artificial intelligence technology for search and KM – a robust, dynamic taxonomy is the engine that will power that approach. The phrase “rubbish in, rubbish out” is widely quoted in data extraction-related contexts. It holds true here too. After all, the artificially intelligent robot will be trained to interrogate just the taxonomy in question to deliver search results. A well-designed taxonomy alongside superior quality content will ensure that the robot in the KM system delivers the right information – which in turn will deliver value and lead to successful KM and AI adoption.
About the author:
With over 25 years of experience in the technology sector, as Managing Director of Lexsoft, Carlos García-Egocheaga is responsible for driving the strategic direction and expansion of the overall business globally. He oversees all aspects of Lexsoft including the P&L, HR, legal and business development.