Paste your tracking plan and an observed event export above, then select Find taxonomy drift to see rogue events, missing required properties, naming drift, and owner cleanup rows.
About the event taxonomy drift card
The event taxonomy drift card compares your tracking plan - the events you intend to track, their required properties, and who owns them - against an observed event export of what is actually firing, and tells you exactly where reality has drifted from the plan. A renamed event, a required property that quietly stopped arriving, a rogue event no one planned, or an event that is no longer firing can all erode trust in dashboards and experiments before anyone notices. This tool surfaces those gaps so a person can route each one to its owner.
Paste your tracking plan on the left and your observed events on the right, select Find taxonomy drift, and the tool matches events by name, flags renames where only the spelling differs, checks each observed event for its required properties, and lists every gap with a severity, an owner, and a plain-language cleanup action. You can download a drift report CSV or copy a markdown cleanup note. Everything runs in your browser. The plan, the export, the event and property names, and the volumes you paste are never uploaded or stored.
How to use
- Paste your tracking plan into the left box as CSV with event, property, required, and owner columns. A sample plan is loaded so you can see how it works.
- Paste your observed event export into the right box as CSV with event, property, and optional count columns. A sample export with a renamed event, a missing required property, and a rogue event is loaded by default.
- Select Find taxonomy drift to match events between the plan and the export and detect the drift between them.
- Review the drift table: each row shows the event, the property, the kind of drift, a severity, the owning team, and a cleanup action.
- Select Download drift report CSV or Copy cleanup note to route the cleanup to the owning teams before dashboards rely on the data.
Worked examples
rage_click flagged Rogue event (owner unassigned)
An event firing in production that is not in the tracking plan is flagged as rogue so a team can adopt it into the plan or stop emitting it.
checkout_completed missing revenue flagged Breaking
A required property declared in the plan but absent from the observed event is flagged breaking because revenue dashboards and experiments depend on it.
Signup Completed flagged Naming drift (plan: signup_completed)
An event firing under a different spelling of a planned name is flagged as a rename, not a rogue event, so the implementation or the plan can be aligned.
Frequently asked questions
- What is event taxonomy drift?
- Event taxonomy drift is when the analytics events actually firing diverge from the tracking plan that documents them: an event gets renamed, a required property stops arriving, a brand-new event appears that no one planned, or a planned event quietly disappears. Any of these can silently break a dashboard, a funnel, or an experiment that assumed the documented taxonomy, so it helps to compare what is firing against the plan before relying on the data.
- How does it match events between the plan and the export?
- It matches events by a normalized form of their name - lowercased with separators and casing removed - so signup_completed, Signup Completed, and signup-completed are treated as the same event. When an observed event matches a planned event only after normalization, it is flagged as naming drift rather than as a rogue plus a missing event. An observed event with no normalized match anywhere in the plan is a rogue event, and a planned event with no match in the export is a missing event.
- How does it decide a property is missing?
- For every event that appears in both the plan and the export, it checks the properties the plan marks as required. If a required property is not present on the observed event, it is reported as a breaking missing property. Properties the plan marks optional are not flagged when absent. Properties that are observed but never declared in the plan are reported separately as unexpected properties at an informational severity.
- What do the severities mean?
- Breaking covers rogue events and missing required properties, which are the most likely to corrupt a dashboard or an experiment. Warning covers naming drift and missing events, which usually need an implementation or plan update. Info covers unexpected properties you can document or remove at leisure. The report lists breaking rows first and carries the owning team on every row so cleanup can be routed quickly.
- Is my tracking plan or event data uploaded anywhere?
- No. The parsing, matching, and drift detection all run in your browser. The plan, the observed export, the individual event and property names, the owners, and the volumes you paste are never sent to a server or saved. Download the drift report or copy the cleanup note before you close the tab.
Use this again tomorrow
Save this page so it's one tap away when you need a quick result.
Ready for a quick Daily Challenge?
Play Daily Challenge on sts.games