TrackLead
←All articles

The Complete Guide to Setting Up Server-Side Google Tag Manager

4 min read

Setting up server-side Google Tag Manager (sGTM) is one of those tasks that sounds simple in a sales deck and turns into an afternoon of debugging in practice. This guide walks through what you actually need, in order, from first access request to the check you should never skip afterward.

This is part of our series for tracking and data specialists. See the full overview here: Server-Side Tracking for Tracking & Data Specialists.

New to the difference between client-side and server-side tracking? Start with Client-Side vs Server-Side Tracking, Explained.

What you need before you start

Get these four things sorted before you configure anything.

1. Admin access to Google Tag Manager

Without publish rights on both the existing web container and permission to create a new server container, you cannot move past step one. Confirm this before you plan anything else.

2. Access to every platform the client sends data to

In practice this means Google Ads, Google Analytics (GA4), Google Tag Manager itself, Meta Business Manager, LinkedIn Campaign Manager, TikTok Ads Manager, and any ecommerce platform involved, such as Shopify or WooCommerce.

Chasing down access to five or six separate accounts, from five or six separate logins, is often the slowest part of the entire project. It is worth requesting all of it on day one rather than discovering a missing login halfway through setup.

3. A hosting environment for the server container

This is where the container actually runs. Some agencies deploy directly to Google Cloud Run and manage the infrastructure themselves.

Most use a dedicated sGTM hosting provider instead, such as Stape, TAGGRS, Addingwell, or AdPage. These handle the server infrastructure so you are not managing raw cloud resources for every client.

4. A clear picture of what is already firing

Look at the client's existing web container before you touch anything. A server-side setup that ignores what is already tracking on the site tends to create duplicate or conflicting data.

Step 1: Create the server container

In Google Tag Manager, create a new server container alongside the existing web container. This is where your server-side tags, triggers, and variables will live.

At this stage, the server container is empty. It does not do anything until you deploy it and start connecting your platforms.

Step 2: Deploy the container to a hosting environment

The server container needs somewhere to actually run. Point it at Google Cloud Run directly, or deploy it through a hosting provider such as Stape, TAGGRS, Addingwell, or AdPage.

A hosting provider typically simplifies this step and adds extra tooling on top of the raw infrastructure, such as dashboards, templates, or EU-based hosting, depending on the provider. Whichever route you choose, you end up with a server URL that your website will send data to.

Step 3: Point your web container at the server

In the web container, update the Google Tag settings (or your GA4 configuration tag) to send data to your new server URL instead of directly to Google. This is the step that actually reroutes traffic through your server-side setup.

Step 4: Configure each destination platform

For every platform on your list, GA4, Meta, Google Ads, LinkedIn, TikTok, and so on, you need a corresponding client and tag inside the server container that knows how to receive and forward that data correctly.

This is usually the most time-consuming step. Each platform has its own requirements, its own naming conventions, and its own quirks.

Step 5: Test before you consider it live

Use GTM's preview mode alongside each platform's own debugging tools, such as GA4 DebugView or Meta's Events Manager test tool, to confirm that events are arriving correctly. Check that the parameters are right and that nothing is duplicating against any client-side tags still firing.

Do not publish a container you have not previewed.

Step 6: Check on it regularly, not just once

This is the step most guides leave out, and it is the one that determines whether a setup still works six months from now. Platforms update their APIs on their own schedule, not yours, and a working setup can silently stop working the moment one of them does.

Put a recurring check on the calendar. Treat "it worked at launch" as a starting point, not a guarantee.

The most common mistakes

Two mistakes show up again and again. The first is forgetting to remove or adjust client-side tags that now duplicate what the server is sending.

The second is treating step six as optional instead of as the step that actually protects the client's data long term.

Where the manual work actually goes

Every step above is real work, and it is the same real work for every new client: chasing down access across five or six platforms, building the container, connecting each destination, testing it, and then remembering to check it again next month.

TrackLead automates the parts that do not need a human doing them by hand. Onboarding is simplified so getting access to a client's accounts is a single guided step instead of a chain of separate invites.

The server container setup itself, creation, hosting deployment, and connections to GA4, Google Ads, Meta, LinkedIn, TikTok, Shopify, and WooCommerce, happens in minutes instead of a half day. Monitoring runs automatically afterward, so step six happens without anyone having to remember to do it manually.