Conversion tracking begins with the outcome
List the actions that matter: a successful enquiry, a completed purchase or a confirmed booking. Separate intent signals from completed results. Clicking a WhatsApp button does not prove that a conversation took place, and opening a form does not prove submission.
Choose a verifiable success point. For a contact form, send the lead event only after successful acceptance. For purchases, use confirmed order information rather than any visit to a thank-you page. Record these definitions so marketing and sales use the same language.
Audit existing tags before installing more
Check the website, ecommerce integrations and Google Tag Manager for existing setups. The same configuration installed directly and through a container can send duplicate events. Document account identifiers, data destinations and which tool owns each event.
Google Tag and Meta Pixel are different systems. They do not automatically report identical totals. On a site with client-side navigation, check subsequent page views and actions as well as the first load; a working initial page is not evidence that the entire journey is measured.
Configure Google and Meta events deliberately
In GA4, mark important events as key events and set up Google Ads measurement through the appropriate linked configuration. If the same outcome is measured directly in Google Ads and imported from GA4, review primary and secondary conversion actions to avoid optimising towards it twice.
For Meta, use an appropriate standard event where available and inspect it with Test Events. If both the browser and Conversions API send the same event, configure deduplication using matching event_name and event_id values. Server-side delivery does not remove privacy or consent obligations.
Test success, failure and repetition
Use Tag Assistant, GA4 DebugView and Meta’s event-testing tools as appropriate. Run a complete journey with test data and verify the event name, timing and relevant parameters. Then attempt the cases most likely to inflate or suppress results.
Include invalid submissions, a confirmation-page reload, navigation between pages and declined tracking where a consent mechanism applies.
- Rejected forms must not count as successful leads.
- One order must not create additional purchases on reload.
- Order identifiers, values and currencies should match the transaction.
- Do not put raw emails or phone numbers into analytics parameters or page URLs.
Review consent and attribution separately
Determine applicable privacy requirements for the audience and data involved, with appropriate advice. Configure tools to respect visitor choices and applicable rules. Google Consent Mode adjusts tag behaviour based on consent signals; it does not collect consent for you.
Differences between platforms can reflect attribution windows, time zones, models and blocked tracking. Align the outcome definition and reporting period before comparing totals. Investigate implementation errors before interpreting every discrepancy as campaign underperformance.
Connect reporting to lead quality
Conversion totals are not a substitute for sales feedback. Record which enquiries qualify and which become customers. Define a suitable way to link outcomes back to campaigns without sending unnecessary personal information to analytics systems.
Efada’s Analytics & Tracking service can help frame a focused measurement review. Start with one important action, establish what the current tools actually record and use that evidence to decide which configuration needs attention.
