Choose the question before the tool
Do you need to know whether visitors find contact, whether the form works or whether enquiries fit the offer? These are different questions. A button click can inform the first but does not prove the others.
Consider a campaign for an initial advisory conversation. The team may distinguish interest, received messages and enquiries judged relevant after review. Write that distinction before configuring events, otherwise the available tool risks defining what the business means by a result.
Define observable conditions
Record what must happen, where it can be verified and who owns it. A thank-you message does not prove receipt when delivery has not been confirmed. Repeated registrations of one step must not be presented as multiple new customers.
Analytics choices, advertising connections, privacy settings and necessary approvals belong in the actual implementation. This guide is decision support rather than a finished technical or legal configuration.
| Event | Definition | Verification |
|---|---|---|
| Contact click | The intended contact route opens | Tested interface step |
| Received enquiry | Message exists in the agreed receiving system | Confirmed test receipt |
| Relevant enquiry | The team judges it suitable | Agreed review process |
| Sale | A defined later business outcome | Appropriate business source, not inference from clicks |
Try success and failure paths
Send fictional material in a safe environment and compare the outcome with the definition. Try a failed submission where the environment supports it. It should not count as a received enquiry simply because a button was clicked.
Check whether reloads or repeated interactions record one event more than once. The technical owner needs to handle the real connection and duplicate protection. Mark untested parts as outstanding rather than as a working integration. Keep the test record so future changes can be assessed under the same conditions.
Give the report a recipient
Agree who reads results, which period applies and which population is included. Small samples and changes to traffic or the offer can complicate comparisons. Report observed outcomes and their conditions rather than promising improvements. The plan becomes useful when someone applies it to a real question and distinguishes technical page behaviour from the business’s later assessment. A carefully defined metric supports decisions; it is not a sales guarantee.