---
title: Your Tags Are Blocked Before They Fire. Tag Gateway Fixes It Before You Build a Server Container.
description: Ad blockers quietly drop a share of your conversion tags. Google Tag Gateway recovers signal through your existing CDN in an afternoon, and it tells you whether server-side GTM is worth building.
author: LETSGROW Dev Team
date: 2026-10-06
category: Analytics
tags: ["Google Tag Gateway", "Server-Side GTM", "First-Party Tracking", "Conversion Tracking", "Consent Mode"]
url: "https://letsgrow.dev/blog/google-tag-gateway-vs-server-side-gtm-first-party-tagging"
---
Your ad blocker problem is a measurement problem you can partly fix in an afternoon, and most teams are paying for a harder fix. Browsers and extensions block third-party tag requests, so a share of your conversions never reaches Google Ads or GA4. The usual answer is server-side GTM, a container you host and maintain. For many teams the better first move is Google Tag Gateway, which serves your Google tag from your own domain through the CDN you already run.

My position: start with the gateway, measure the recovery, and only build a server container when you have a specific job the gateway cannot do.

## What Each Option Actually Does

Google Tag Gateway routes requests for the Google tag and its measurement calls through a path on your own domain, such as yoursite.com/metrics, instead of googletagmanager.com. Your CDN or load balancer proxies those requests to Google. The browser sees first-party traffic, so filter lists that target Google's domains have less to match against. Setup is a CDN integration plus a toggle in the Google tag settings, and Cloudflare, Akamai, Fastly, and Google Cloud load balancers are supported paths.

Server-side GTM is a different animal. It is a separate container running on infrastructure you own, typically Cloud Run or App Engine. The browser sends one hit to your endpoint, and the server container fans it out to GA4, Google Ads, Meta, and anywhere else. You get control over the payload, which means you can strip fields, enrich events, and enforce consent rules before data leaves your perimeter.

The gateway is a pipe. The server container is a processor. Those are different tools, and picking the processor when you needed the pipe is how teams end up with a monthly hosting bill and a container nobody monitors.

::compare-table
| Dimension | Google Tag Gateway | Server-Side GTM |
| --- | --- | --- |
| Setup effort | A CDN integration and a settings toggle | A new container, hosting, a custom domain, and client and tag configuration |
| Ongoing cost | Usually your existing CDN plan | Cloud hosting plus engineering time to patch and monitor |
| What it fixes | Tag requests blocked because they come from Google domains | Blocking, payload control, and fan-out to many vendors |
| Data control | Limited, requests pass through to Google | Full: filter, hash, enrich, and drop fields before forwarding |
| Failure mode | Misrouted path breaks the tag | Container outage drops events for every destination |
| Best fit | Teams that mostly run Google tags | Teams sending data to many platforms with strict governance |
::

## Where Teams Go Wrong With Server-Side

The common failure is treating server-side as a recovery tool when it is a control tool. Moving a tag from the browser to a server does not make it first-party by itself. If your server container is served from a subdomain you control, you get the cookie and blocking benefits. If it runs on a default cloud URL, you traded one third-party domain for another and paid for the privilege.

The second failure is ownership. A server container is production infrastructure. It needs a custom domain, TLS, autoscaling limits, logging, alerts, and someone who gets paged when it stops forwarding. Marketing teams that spin one up through a quick-start wizard often discover months later that it ran at a single instance and dropped events during a traffic spike. Nothing errored in the browser, so nobody noticed.

The third failure is consent. Server-side does not exempt you from consent. Your Consent Mode signals still have to travel with the event, and the server container has to respect them. If you strip the consent state while forwarding, you have built a compliance gap with better branding.

## A Rollout You Can Defend

Run this in order, and stop when the numbers stop justifying the next step.

First, baseline. Pull seven days of conversions from your CRM or order system and compare them with what Google Ads and GA4 report for the same window. That gap is your measurement loss, and it is the number every later step is judged against.

Second, enable Tag Gateway on one property. Keep the same CDN, add the first-party path, and verify in the network tab that the tag loads from your domain. Wait two full weeks so you cover a normal business cycle, then compare the gap again.

Third, decide whether a server container earns its keep. The honest triggers are: you forward events to four or more platforms and want one source of truth, you need to remove personal data before it leaves your control, or you need to enrich events with server-side data such as order margin. If none of those apply, you are done.

::checklist
title: Before you build a server container
items:
  - Measured the gap between CRM conversions and ad platform conversions for at least seven days
  - Enabled Tag Gateway and re-measured the gap after two weeks
  - Named a specific job the gateway cannot do, such as multi-vendor fan-out or field-level redaction
  - Assigned an owner who is paged when the container stops forwarding events
  - Mapped the container to a first-party subdomain with TLS and autoscaling limits set
  - Confirmed Consent Mode signals are passed through and enforced on the server
  - Set an alert on event volume so a silent drop is caught within a day
::

## The Takeaway

Do not let a vendor webinar decide your tagging architecture. The gateway costs close to nothing and tells you how much signal you were losing. If the recovered number is small, you saved yourself a hosting bill and an on-call burden. If it is large and you still have governance needs, you now have evidence to justify the server container and a baseline to prove it worked.

This week, run the baseline comparison and check whether your CDN is on the supported list. That is two hours of work, and it settles the question with your own data.
