How can you create effective taxonomies for digital products?

The best way to understand your audience is to visualize and empathize with them. To achieve this I would typically create a User Experience Map to understand where in the journey they encounter the most friction. Only in those moments of friction will you be able to understand what the user is feeling and what they want to achieve. Additionally you should build out Personas for each unique type of users who would be engaging with your product. Focusing mainly on their goals, motivations and pain points. These frameworks should be top of mind when building out your IA, content hierarchy and UX design in general.

A taxonomy can be used for structure or attribution; or both. Structure resembles a categorization (structure for the good of users), attribution for the good of information. If you want a sustainable taxonomy which can be used over a long period of time you'd need to go for the latter (see Linnean taxonomy). An organization or process centric taxonomy will only last until your organization or process changes (3-7 years) until you have to redo from scratch. Better then to build a proper taxonomy with attribution, then apply a categorization on top for specific purpose.

Taxonomies serve as master data, facilitating the mapping, categorization, search, and explanation of digital products. These use-cases directly enhance user interaction with the content and functionalities offered by the products. Given the high priority digital products place on customer experience and journey, taxonomies play a crucial role. They are captured and maintained by the product team to ensure that users can easily navigate and engage with the products, ultimately enhancing their overall experience.

By definition, taxonomies are hierarchical. Taxonomies can also be polyhierarchical, although the complexities of a poly hierarchy can be difficult to govern and maintain.

It's more appropriate to model what is called a “networked” taxonomy as an ontology, particularly if your project is about Linked Data and AI -related techniques like RAG. An ontology captures the semantics and complexity of a "networked" taxonomy, while also providing more benefits such as consistency, interoperability, reasoning, and inference.

Following the ANSI Z39.19 standard is advised as the nuances of a well structured taxonomy follows a rule base, designed to support information discovery and transmission .

We're committed to web taxonomies now, individuals not trained in taxonomy will not succeed in creating a taxonomy. They will create categories and this is where card sorting and similar techniques come into play, but this should not be confused with creation of a taxonomy.

In digital product development, there's a temptation to rely heavily on expanding taxonomies, which can lead to errors in data-driven functionalities. For instance, a platform categorizing documents based on taxonomies may misclassify important ones if not actively managed. To mitigate risks, product teams must balance taxonomy expansion with effective management. Regular reviews and collaboration with operations teams are vital to ensure taxonomies accurately reflect user needs and business objectives. This approach ensures a seamless user experience while leveraging data-driven functionalities.