Server-side tracking moves event capture from the visitor's browser to your own server. That sounds dry, but in practice it means 20-40% of conversions recovered that Safari, Brave, and ad-blockers eat today.
Client vs. server
Client-side: GA4 / Meta Pixel run directly in the visitor's browser. Ad-block filters see them, Safari ITP deletes their cookies, and iCloud Private Relay hides the IP address. The result: a dataset with a leaky attribution model.
Server-side: the browser sends the event to your own server
(typically track.yourstore.cz), which forwards it to GA4, Meta
CAPI, and Google Ads. Ad-blockers miss it — it looks like a first-party
request. ITP has nothing to delete — we issue the cookies, not a third party.
Why it's a necessity now
- iOS 17 + iCloud Private Relay: anonymizes the IP, confuses attribution for Meta and Google Ads.
- Safari ITP 2.3+: deletes _ga, _gid, and _fbp cookies within 7 days.
- Brave + uBlock: block 100% of requests to
analytics.google.com. - Consent Mode v2 in the EU: without a server-side model you have no way to capture "denied" events without PII.
What you gain
Real numbers from our onboardings: a store doing CZK 1M/month on Meta Ads saw +34% conversion match rate in the Events Manager after switching to server-side. That means CAPI got more data → the algorithm knows more → ROAS grows by 15-25% with no budget increase.
If you want a rough estimate for your case, use the ROAS calculator — it computes how much revenue you're missing due to lost tracking.