Scalable Taxonomies

In working on enterprise taxonomy projects, the issue of taxonomy scale always comes up. “Scale” is a word that gets tossed around a lot but what does it really mean for taxonomies? How can taxonomies scale, and what do organizations need to think about when considering requirements for taxonomy scale?

We have found that the best way to define taxonomy scale is to break it into three sets of considerations. They are as follows: 

Taxonomy Size

The ability to grow the taxonomy by adding new terms or relationships between existing terms. Examples here include growing a topic taxonomy to easily incorporate new terms identified in search log or the ability to support the addition of new high-level categories. This later example would be essential for a company that is planning on growth by adding new product lines or via acquisition.  This type of growth may have significantly different requirements than a taxonomy that will need to scale to support the addition of synonyms or more sophisticated relationships.  In the former case, the taxonomy and governance plan will need to be tightly tied to the capabilities of the systems or analytics that are dependent on the taxonomy.  In the later case the taxonomy and the taxonomy management system would need to support a high velocity workflow.

Taxonomy Function

The ability to grow a taxonomy to support new (known or unknown) use cases is another common scenario where a taxonomy may need to support the scaling of the overall business landscape.  Planning for the growth of use cases can be an extremely challenging scenario as it requires that a lot of thought and planning go into the overall model.  A well-designed model that takes into account the full set of requirements (current and anticipated) and modeling best practices is really a thing of beauty.  Some of the most satisfying situations a taxonomist can find herself in are when a taxonomy model has been created and people start asking about how it will address different scenarios or use cases (new queries it needs to support, usage by additional systems, multi-lingual requirements, new product categories, etc). The successful model will have anticipated these requirements and be able to support them with minimal adoption cost.

No model can be all things to all people, but this is a case where the taxonomist needs to really understand the business environment where the taxonomy will be used AND do the extra work to make sure that the model follows best practices. For example, there is always a point in a project when there is a random collection of terms that only seem to fall under some notion of “miscellaneous”.  The work required to really find the best place for them in the taxonomy may seem daunting.  However, taking a step back and figuring out how to address them is almost always worth the extra effort.  This will likely require the organization to take a deep look at why those terms are there to begin with (the number of times we have had to find a place for branded beer cozy’s …)  And this ability to look around corners and know where to abandon taxonomy standards in favor of supporting business requirements or where to push back on the business requirements to support a rational model is the crux of a successful taxonomy.

Taxonomy Usage

The ability to support the addition of new products is often a requirement that comes up when considering the requirements of a navigation or product category hierarchy taxonomy.  While the products themselves will not be part of the taxonomy the taxonomy or navigation structure will often need to change to support the new products, even when the products are all associated with existing product categories.  In this case “scaling the taxonomy” is really a process of updating it to make sure it is well balanced and in terms of product distribution.  For example if a retailer increases the number of items in a category, they may likely need to increase the number of sub categories to support the end user experience or the needs of their merchandisers. While this may seem to be the easiest for the taxonomist to manage, it is often the scalability scenario that companies are most concerned about.  Taking the time to make sure that taxonomies can be adjusted accordingly will be essential to support this type of scalability requirement.

Scalability ultimately comes down to creating a model that can be managed in the technology stack and designing a governance process that will allow the growth to occur in a way that is manageable and durable over time. The better the taxonomist’s understanding of the business landscape and business requirements the better they will be able to understand what type of scalability is required and the more likely they will be able to create a robust model that will withstand the test of a hundred new use cases, marketing initiatives, or product lines.