Event taxonomy

An event taxonomy is the agreed set of names, properties, and structure used to record what users do in a product — for example, deciding that a purchase is always logged as an order_completed event with consistent properties like order_id, value, and currency, rather than being named and structured differently by every team that instruments it.

It is a naming and modeling convention, not a data collection mechanism itself: a well-designed taxonomy typically distinguishes objects (what was acted on) from actions (what was done), keeps naming consistent (verb tense, casing), and defines which user properties attach to which events. It differs from a tracking plan, which is the operational document listing every event to implement, who owns it, and when it fires — the taxonomy is the rulebook the tracking plan follows.

An event taxonomy matters because without one, the same user action ends up logged under different names by different teams or platforms, breaking cross-feature analysis, path analysis, and any metric that spans more than one instrumented area. The most common pitfall is inconsistency introduced over time as new features ship: names drift, duplicate events appear, and properties are populated differently on web versus mobile, which is why taxonomies need active governance, not just an initial document.