Analytics installed after launch means every early visitor — often the highest-intent traffic, from a launch announcement or press mention — goes completely unmeasured. There's no way to retroactively recover that data once the moment has passed. Setting up tracking is a pre-launch task, not a "get to it in week one" task.
GA4 setup checklist
- A GA4 property is created and a data stream is connected to the live domain (not a staging or development URL).
- The measurement ID is installed on every page of the site, not just the homepage — a partial install is one of the most common tracking gaps, especially on sites with pages built outside the main template.
- Tracking is verified live using GA4's Realtime report or DebugView while browsing the site in a separate tab, confirming pageviews and events actually register.
- Internal traffic (the site owner's own visits, and any agency/dev team testing) is filtered out using an internal traffic filter, so launch-week analytics aren't skewed by the people building the site.
- Cross-domain tracking is configured if checkout, booking, or any part of the funnel happens on a separate domain or subdomain.
Tag manager checklist
- The tag manager container (Google Tag Manager or equivalent) is installed in both the
<head>and immediately after the opening<body>tag, per the platform's own installation instructions — skipping the body snippet causes it to work inconsistently for users with slow connections. - Tags fire correctly across every page template, not just the one used during setup and testing.
- No duplicate tags exist from a previous platform or template — duplicate GA4 tags double-count every pageview and event.
- Tag firing is verified using the tag manager's own preview/debug mode before publishing the container live.
Conversion event checklist
- The 3-5 actions that actually matter for the business — form submit, purchase, signup, key page view — are defined explicitly as conversion events, not left as generic pageviews.
- Each conversion event is tested manually to confirm it fires exactly once per real action, not zero times or multiple times.
- Ecommerce tracking (if applicable) captures transaction value, not just a binary "purchase happened" event — value data is what makes ROI reporting possible later.
- Conversion events are marked as such inside GA4's own configuration, not just sent as regular events, so they surface correctly in reporting and any connected ad platforms.
| Check | How to Verify | Common Failure |
|---|---|---|
| GA4 installed sitewide | Check Realtime report while browsing multiple page templates | Measurement ID missing from pages built outside the main template |
| Tag manager firing correctly | Use the platform's preview/debug mode before publishing | Container published without ever being tested in preview mode |
| Conversion events accurate | Trigger each event manually, confirm it appears once | Event fires on every page load instead of only on the real action |
| Internal traffic filtered | Check that the dev/agency team's own visits are excluded | Launch-week data skewed by the people building the site |
This setup work overlaps directly with a broader conversion tracking validation pass — the difference at launch is timing: none of this can be "fixed later" for the traffic that already happened. Get the events right before the first real visitor arrives.
FAQ
How do I confirm GA4 is actually tracking correctly?
Open GA4's Realtime report, browse the live site in a separate tab, and confirm your own visit and any test events (form submits, button clicks) appear within a minute or two — DebugView gives an even more detailed, event-by-event view for confirming parameters are captured correctly.
- Realtime confirms basic tracking is alive; DebugView confirms individual event parameters are being captured correctly, which matters for conversion accuracy.
- Test from a separate device or incognito window so your own regular browsing habits don't muddy the test.
What conversion events should be set up before launch, not after?
The 3-5 actions that actually indicate business value — typically a form submission, a purchase or checkout completion, an account signup, and any key page view like a pricing page — should be defined and tested before traffic arrives, because GA4 can't retroactively reconstruct conversion data for traffic it wasn't configured to measure.
- Picking too many conversion events dilutes reporting; 3-5 genuinely decision-relevant actions is usually the right range.
- Ecommerce events should capture transaction value, not just a yes/no purchase flag, to make ROI reporting possible.