Tracking & Analytics · Server-Side Tracking

Server-Side Tracking with Your Own sGTM

A tagging server on your own subdomain processes events before ad blockers and browser restrictions filter them out. We build sGTM pipelines, run them, and tell you upfront what they cost and what they deliver.

01 — Foundation

What an sGTM is and why first-party wins

We have been building tracking setups since Google Analytics existed, and the shift to server-side is the biggest architecture change since the Urchin days. A server-side Google Tag Manager is a second GTM container that runs not in the browser but on a server under your control. The browser sends just a single event stream to a subdomain like track.yourdomain.com, and from there the server distributes the data to GA4, Meta, Google Ads and every other endpoint.

The decisive piece is the first-party subdomain. Ad blockers filter requests to known tracking domains, and Safari's ITP shortens the lifetime of cookies set by third-party servers. A subdomain of your own, with its own DNS record, is part of your website as far as the browser is concerned. Requests get through, cookies live longer, and the gap between real orders and measured conversions shrinks noticeably. How this fits into the bigger picture is covered on our Tracking & Analytics overview.

02 — Current

What's changing in server-side tagging right now

Google is visibly expanding the infrastructure. The Google Tag Gateway for Advertisers, a leaner flavor of first-party routing, is now generally available in its Google Cloud integration, and since this year there is a guided workflow to set it up via Amazon CloudFront as the CDN. For some setups that is enough, but a full sGTM stays more flexible because it serves every vendor and can transform events.

Inside sGTM itself, version 3.2.0 cleaned house: the GA4 client no longer loads gtag.js, and all Google libraries now come through the web container client. On top of that, Google is improving attribution for Google Ads-linked Analytics properties via parallel browser signals when Google Signals is active. Changes like these roll out quietly in the background, but they can shift existing setups. That is exactly why monitoring the release notes is part of how we run these systems.

Our take: Tag Gateway is a good entry point, but it does not replace your own tagging server. If you want to enrich, filter or distribute events to multiple platforms, you still need the full sGTM.

03 — Operations

Hosting options, honestly compared

An sGTM needs a server, and there are three sensible routes. Google Cloud Run is the default: it scales automatically, typically lands between 50 and 150 euros a month for small and mid-sized shops, but requires someone to keep an eye on instances, updates and logs. AWS works just as well, though it really only pays off if your infrastructure lives there anyway. Stape is the managed route: entry from around 20 euros a month, operations and updates included, but less control and tiered pricing that climbs with heavy traffic.

Our honest assessment: for most clients without their own DevOps team, Stape is the sensible start. Beyond a certain volume, or with specific data residency requirements, moving to your own cloud infrastructure pays off. We run those numbers before the setup, not after. If you are unsure where you stand, we will look at your existing setup in a free short audit, just write to us below.

GCP Cloud Run AWS Stape Cost Monitoring Uptime Checks

04 — Engineering

Event enrichment and cookie lifetime

The real payoff of a tagging server is that events can be refined there. The server can validate order values against the shop backend, add margin data, attach click IDs or discard bot traffic before anything reaches the platforms. It also sets cookies as genuine HTTP cookies from your own domain, which ITP does not cap after seven days. Returning buyers stay recognizable as returning. These enriched events are also the foundation for the Conversion APIs of Meta, Google and friends.

05 — Privacy

Data minimization as a built-in advantage

Client-side, every request goes to the vendors unfiltered, IP address and everything the tag picks up included. Server-side, you decide per endpoint which fields go out: truncate or drop the IP, reduce user agents, strip internal parameters entirely. That is data minimization in practice, and it makes conversations with your data protection officer noticeably more relaxed. Proper consent remains the prerequisite, which is why we always build the pipeline together with Consent Mode.

Server-side does not mean collecting more data. It means deciding for yourself which data flows where.

06 — Practice

Migration from the pure web container

Hardly anyone starts from scratch. Our typical migration runs in stages: first the subdomain and the server, then GA4 through the new pipeline in parallel with the old setup, with the numbers reconciled against each other. Only when both pipelines measure the same thing do the ad platforms move over and the old tags come out. That way there is no blind flight in the middle of a quarter. You notice the result directly in the account, by the way: a better data foundation means better Smart Bidding, which we use daily in Google Ads.

07 — Let's talk

How much does your tracking really measure?

Tell us briefly what you currently run and how many orders or leads flow through per month. We will give you an honest estimate of whether a tagging server pays off for you. First call free, 30 minutes.

We reply within one business day.