In this article
  1. What should conversion measurement testing verify?
  2. What should the event specification say?
  3. Which test cases reveal misleading counts?
  4. Where should the evidence be inspected?
  5. What information should stay out of analytics?
  6. What should the completed QA record contain?

What should the event specification say?

Write the business meaning, trigger, allowed parameters and exclusions before configuring tags. Identify the source of truth for the underlying action. The specification should say whether a page reload, repeated click or existing-customer message may count again.

Google's recommended event documentation distinguishes lead generation from later qualification and outcome events. Use an appropriate name when it fits, but make sure the implemented condition reflects the intended meaning. Naming an event generate_lead does not validate the form or qualify the person.

Which test cases reveal misleading counts?

Which test cases reveal misleading counts?
CaseExpected checkBusiness evidence
Valid requestThe defined success event appears as intendedAccepted request and correct receiving record.
Invalid required fieldNo completed-submission event is claimedValidation prevents acceptance.
Destination failureThe event does not conceal the failed taskFailure record and customer-facing message.
Repeat click or reloadCounting follows the specified duplicate ruleNo unintended second business action.
Different consent stateCollection follows the approved behaviorDocumented state and observed event behavior.

Where should the evidence be inspected?

Use the browser or tag-testing view to see the trigger, then inspect the collected event and its parameters. Google's DebugView guidance explains how to inspect events during testing and notes that privacy controls can affect visibility. Record the test conditions so missing data can be interpreted correctly.

Check the receiving inbox, booking tool or CRM independently. Analytics collection may be limited while the business request still succeeds; the reverse can also occur. A useful QA record explains that difference without bypassing the approved collection controls.

What information should stay out of analytics?

Use a non-personal test reference where a reference is needed. Avoid names, email addresses, phone numbers and free-text messages in ordinary analytics parameters, URLs or titles. Google's PII guidance specifically addresses these accidental collection paths.

Inspect the values actually sent, not just parameter labels. A harmless-looking field can contain information entered by a user. Test a safe sample and review the resulting payload before deploying a configuration that copies input into reports.

What should the completed QA record contain?

Record the URL, date, environment, collection state, action, expected event and observed receiving result. Include any unexplained discrepancy and its owner. A checklist is useful when another reviewer can reproduce the conclusion without guessing which version was tested.

The GA4 setup checklist covers the wider configuration. For implementation help, conversion tracking services begin with the required $300 Growth Audit, with the selected work quoted separately by AI Growth Systems.