Prove it moved the number, not that it ran green

Here is the failure mode that catches nearly everyone who sets this up: they see a 200 response, watch an event land in GA4, declare victory and move on. A 200 means the event sent. It does not mean your attribution improved, your cookies now persist, or your conversions reconcile. Green means it ran, not that it works. The entire value of server-side tracking lives in numbers you have to deliberately go and check, so you assert the value, not the presence.

The verification checklist

Three checks, each of which would fail if the relocation had not actually delivered its benefit.

First, confirm events arrive in GA4 via the server, not via the old client-side path running in parallel. Open GA4 DebugView and trace an event through. You are not checking that data arrived; you are checking it arrived through your tagging subdomain. If your web container is still firing GA4 directly to Google alongside the server, you have two pipes and have proven nothing about the new one.

Second, and this is the one almost nobody runs, check the Safari cookie now persists past 7 days. Open your site in Safari, complete a visit, then inspect the relevant first-party cookie and read its expiry. A server-set cookie should show an expiry well beyond a week. If it still expires in 7 days, or 24 hours on an ad-click landing, your cookie is still being set client-side and the central benefit has not landed despite every event returning 200. This single check separates a setup that works from one that merely runs.

Third, reconcile measured conversions against the ad-platform numbers. Pull conversions as counted by your server-side GA4 and compare them to what Google Ads and Meta report. The point is not perfect agreement, it is direction and magnitude: the server-side count should be closing the gap that client-side tracking left open.

Size your own recovery

Do not trust a vendor's headline figure for how much you will gain. Industry case data does report meaningful lifts, with vendors citing increases in measured pageviews and conversions of roughly 5 to 30% from resisting ad blockers and tracking prevention, but that is a range across many businesses, not a promise for yours. Run a proper before-and-after on measured conversions for two to four weeks and read your own number. Your audience's block rate, your traffic's Safari share and your sales-cycle length all move the figure, and the only number that should drive your decisions is the one your own data produced.

Watch the console and the logs

Throughout, read the browser console and the server logs on each path you test. The characteristic failure of a tracking setup is a silently dropped event: no error surfaced to the user, no exception, just an event that quietly never made it. A check that asserts a specific event arrived, with the right parameters, on the right path, is the only thing that catches a silent drop. Presence checks pass straight through it.