Google Analytics Conversion Tracking: Events, Key Events and a Validation Workflow

Define the action before creating an event

Write down the exact action you want to count. For a lead form, specify whether success means a completed server response, a thank-you page visit or a CRM record. A button click can happen before a validation error. Counting that click as a completed lead changes the meaning of every later report.

Use an existing event only when its trigger matches the action. Check the event name, where it fires and which parameters identify the result. Avoid personal information in event names, URLs and parameters. Give your analyst a definition they can compare against the website and your business records.

Keep three checks separate: the event fires, the property marks the intended event as important, and advertising uses the intended conversion action. Google’s distinction between conversions and key events explains why those labels now answer different reporting questions.

Inspect key events and their counting methods

Ask your assistant to list the connected property’s current key events. SourcetoMCP returns event names, counting methods, default values and currencies, plus whether Google permits removal. Keep any truncation warning, because a short list with more rows available is an incomplete audit.

SourcetoMCP can create a key-event designation, update its counting method or default value, and remove a designation when GA4 permits it. It does not install an event tag on your website. Creating a key event with a name your site never sends leaves the collection problem unresolved.

Calculate what repeated events do to the count

Suppose a hypothetical test has ten sessions with a form event. Eight sessions send it once and two sessions send it three times. That produces fourteen event occurrences across ten sessions. This is a labelled example, with no claim about a customer account.

With a once-per-event key-event method, those occurrences can produce fourteen key events. With once per session, they produce ten. Google documents both counting methods. Neither setting tells you whether the two repeated sessions contain genuine additional leads or an incorrectly repeated trigger.

If the same period has 200 sessions, fourteen divided by 200 is 7%, but that is an event-per-session ratio. Ten sessions with the action divided by 200 is a 5% session-based rate in this example. State the numerator and denominator when you show a percentage.

Repair duplicate firing at the collection source when the extra events are errors. Changing the counting method can hide repeated events in one measure while leaving the underlying collection behaviour unchanged.

Validate one successful action in the browser

Use a test device to complete the action from start to finish. Record the time, page, expected event and consent state. Check a failed form attempt too. The failure should follow your documented measurement plan rather than silently count as a successful submission.

GA4 DebugView lets you inspect debug-mode events and parameters. Use Tag Assistant or the website’s tag testing setup to identify the test device. Confirm that the right property receives the event once for the intended action.

SourcetoMCP’s realtime report can show event names and counts from roughly the last thirty minutes. It is an aggregate report and can lag. It cannot replace DebugView, show every event parameter or prove that a particular business record came from the exact event you tested.

For historical checks, the standard performance operation supports selected dimensions and an aggregate metric named conversions. In this connector that alias counts the property’s GA4 key events. It does not expose arbitrary historical event-name breakdowns or Google Ads advertising-conversion reports. Use GA4’s interface for the event-specific reconciliation.

Run a read-only AI review before editing

Audit GA4 conversion tracking for property [ID] without changing anything. List current key events with names, counting methods and default values. Read Google Ads links and a complete-period performance report using supported fields. Explain that the connector’s conversions metric counts GA4 key events. Keep event-level browser validation, historical event-name analysis and Google Ads conversion reconciliation as separate checks. Return evidence, missing information and an exact proposed configuration edit for review.

Compare the report with your written measurement plan and the GA4 audit checklist. Keep missing values distinct from missing events. The not set diagnosis guide covers absent dimension values that can occur even when activity was collected.

If you approve a setting change, name the property, event and exact field values. Read the saved key-event list again afterwards. Treat the read-back as configuration verification and retain the browser evidence as collection verification.

Connect advertising measurement without double counting

Linking GA4 and Google Ads does not prove the right event is used for bidding. Inspect the Google Ads conversion action, the selected goal and any existing website-tag action that measures the same outcome. Two actions can represent one purchase and need deliberate reporting rules.

Google’s conversion creation guidance covers creating advertising conversions from Analytics events. Complete that configuration in the supported Google interface and validate the resulting action. SourcetoMCP’s GA4 key-event write does not perform that import.

Use the Google Ads conversion-tracking audit for the next check. Record the action IDs, attribution timing, counting rules and reporting period before comparing totals. A difference between two totals needs that context before you classify it as a tracking error.