Practical working guide
A website lead measurement plan that separates clicks from customers
Measure a service website with separately defined enquiry, qualification and customer outcomes, each backed by an authoritative record. A button click is an interaction, not proof that an enquiry was received or a sale completed. Define deduplication and reporting limits before comparing conversion rates.
Create an evidence-based measurement dictionary
| Milestone | Details |
|---|---|
| Enquiry received | Evidence: Persisted valid submission reference Do not substitute: Click on Submit |
| Qualified lead | Evidence: Recorded qualification decision Do not substitute: Form completion alone |
| Customer outcome | Evidence: Defined accepted commercial record Do not substitute: Visit to a thank-you page |
| Attribution | Evidence: Known source and declared matching limits Do not substitute: Guess based on an email domain |
Work the denominator before celebrating a rate
Fictional example: 40 form clicks produce 25 stored submissions. Three are duplicate requests and two are spam, leaving 20 unique valid enquiries. Eight are qualified: the enquiry-to-qualified rate is 8/20 = 40%, not 8/40. State the period, cohort and treatment of still-pending cases. If qualification happens later, do not mix today’s enquiries with unrelated decisions from previous months.
Choose events after agreeing the business meaning
GA4 documents recommended events including generate_lead and qualify_lead. Map them to your verified definitions instead of treating the names as proof of an outcome. Record event owner, trigger, deduplication key, test method and permitted data. Keep personal details out of analytics payloads and follow applicable consent requirements. This is a measurement specification, not a claim of new dashboards, automatic attribution or already-installed event tracking.
Sources: Google Analytics — Recommended events.