Server-side tracking
How it works
The term covers two approaches that are often combined:
- Server to server through conversion APIs: an event first arrives on your server or at a tracking service. From there it goes straight to the ad platforms through their Conversion API.
- Server-side tag manager: the browser sends data to your own server container, such as server-side Google Tag Manager, often on a subdomain of your website. The container passes it on to analytics tools and ad platforms.
Either way, a script usually stays in the browser to capture what happens on the page. The server decides which data goes where.
Why it matters
Browser-only tracking loses data. Ad blockers block known tracking domains, Safari limits cookies set by JavaScript, and some pages are closed before the pixel has fired. A server doesn't depend on any of that, and it keeps control over which fields go to which platform. Events without a browser at all, such as payments or deal stages from the CRM, can only be reported server-side.
TSO takes the first approach: the script captures events in the browser, conversions go to the platforms through their conversion APIs, and individual fields can be withheld per destination.
Example
An online course seller takes payments on a payment provider's checkout page. The purchase happens on another domain, and the pixel on the seller's thank-you page only fires if the buyer comes back to it. Of 50 purchases a month, 35 reach the pixel.
Server-side it looks different: the payment provider reports every purchase by webhook. The server matches it to the original visit through the email address and the stored click ID and sends all 50 purchases to Meta and Google Ads.
Common mistakes
- Treating it as a way around consent: server-side tracking doesn't change which consents you need. Clarify the legal side with your data protection counsel.
- Dropping the browser entirely: without a script in the browser you lose the click IDs, cookies and page views a platform uses to match an event. Server and browser complement each other.
- Duplicate events: when pixel and server run side by side, every event needs a shared event ID, or the platform counts it twice.
- Underestimating the upkeep: your own server container has to be run, monitored and kept up to date. If it fails, data goes missing without anyone noticing.