Skip to main content
Card class: Non-HeroCategory: Analytics

At a glance

Event Taxonomy Drift (new / unexpected events) is an event-health metric tracked from Amplitude data. It counts event names that are new or unexpected in the window. A sudden cluster of new event names usually means an untracked release shipped, or an existing event was renamed, which quietly breaks the historical funnels and cohorts that referenced the old name. The card pairs with the other event-health metrics below to keep your tracking trustworthy.
What it countsThe number of event types appearing in the window that are new or were not expected against your established taxonomy, as exposed by the Amplitude integration. Computed from the latest available data and refreshed on the standard data refresh.
Sample typeBackend API data from Amplitude, refreshed on the standard data refresh.
Why it mattersNew event names are how tracking quietly drifts. A renamed event keeps historical funnels pointing at a name that no longer fires, while a new untracked event signals a release that analytics was not told about. Watching drift keeps your reporting aligned with what the store actually does.
Reading the valueA low, stable count is healthy. A spike in new event types in a short window is the signal to investigate which release or rename caused it. Compare the current window to the prior one to judge whether drift is normal churn or a structural change.
Currencycount
Time window7D
Alert trigger>5 new event types
Sentiment keyamp_event_taxonomy_drift
Rolesowner, marketing

Calculation

Calculated automatically from your Amplitude data. The card compares the set of event names seen in the window against your established taxonomy and counts those that are new or unexpected. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.

Worked example

A representative reading of Event Taxonomy Drift (new / unexpected events) for a typical merchant on Amplitude. In the week ending 12 Mar 26 the card reports 1 new event type, well within normal churn. In the week ending 19 Mar 26 it jumps to 7 new event types, crossing the alert threshold. On inspection, a frontend release renamed add_to_cart to cart_add and introduced five experimental events. The rename is the dangerous one: every funnel and cohort built on add_to_cart now misses the new firings, so conversion looks like it collapsed when nothing actually changed for shoppers. Vortex Mind can correlate the drift spike with the deploy window and flag which historical cards reference the old name; ask Ask Viq “which new event names appeared this week and which old events stopped firing” to confirm the rename before trusting any funnel reading.

Sibling cards merchants should reference together

CardWhy merchants reach for it
amp_core_events_firingConfirms whether a rename has silently stopped a core event from firing.
amp_event_volumeShows whether drift coincides with a change in overall ingestion volume.
amp_top_eventsReveals whether a new event has displaced an expected one in the rankings.
amp_alert_tracking_brokenFires when drift turns into a core event going dark.
amp_health_scoreRolls drift and other signals into one overall tracking-health read.

Reconciling against Amplitude

Where to look in Amplitude’s own dashboard: The Data or Govern section is the home of event taxonomy. Use it to see newly ingested event names, blocked events, and naming conventions, and to confirm whether a name is genuinely new or a rename of an existing one. Event Segmentation lets you chart a suspected new event against the one it may have replaced. Confirm the period and any segment filters match the Vortex IQ profile to reconcile cleanly. Why the Vortex IQ value may legitimately differ:
ReasonDirectionWhat to do
Period boundary. Vortex IQ uses a rolling 7-day window by default; Amplitude’s taxonomy views may show all-time or calendar ranges.VariableMatch the period range.
Time zone. Amplitude uses the project time zone; Vortex IQ aligns to the merchant reporting time zone, shifting which day a new event first appears.MarginalConfirm time zone match.
Filter or segment scope. Profile-level filters may exclude events from test or internal traffic that Amplitude’s unfiltered taxonomy still lists.VariableMatch filter settings.
Cross-connector reconciliation: when a rename is suspected, cross-check the new event’s volume against the old event’s drop and against orders in your ecommerce connector to confirm shopper behaviour did not actually change. For divergence investigations, use Vortex Mind.

Known limitations / merchant FAQs

Q: How often does Event Taxonomy Drift (new / unexpected events) update? The card refreshes on the standard data refresh (typically every 30-60 minutes for live integrations). For real-time signals after a release, force a manual refresh from the dashboard. Q: Why does my Amplitude dashboard show a different number? The most common reasons are period-boundary differences (Vortex IQ uses a rolling 7-day window vs Amplitude’s all-time or calendar taxonomy views), time-zone alignment, and filter scope (profile-level vs Amplitude’s unfiltered list). Match these settings before assuming a real divergence. Q: Is every new event a problem? No. Some drift is healthy as you ship features and add tracking. The risk is renames that orphan historical funnels, and untracked releases that add noise. The card surfaces the count so you can decide which new names are intentional and which need cleanup. Q: Can I customise the alert threshold? Yes. The default alert fires above five new event types in the window, but you can tune that and the taxonomy baseline per profile in the Sensitivity tab to match how often your team ships.

Tracked live in Vortex IQ Nerve Centre

Event Taxonomy Drift (new / unexpected events) is one of hundreds of KPI pulses Vortex IQ tracks across Amplitude and 70+ other ecommerce connectors. Nerve Centre runs the detection layer; Vortex Mind investigates the cause when something moves; Ask Viq lets you interrogate any number in plain English. Start for free or book a demo to see this metric running on your own data.