Google Tag Gateway vs Server-Side Tagging: What Each Fixes
See what Google tag gateway and server-side tagging each recover for a store, which pays off at your size, and why a wrong purchase goal gets past both.
Google tag gateway and server-side tagging each recover part of the conversion data lost in the browser: orders you already had, not new sales. The gateway serves the Google tag from your own domain and is free on Cloudflare, so it suits almost any store. A tagging server costs about $90 a month by Google’s own estimate, plus upkeep. It pays off only for a specific need, such as your own conversion value. Both leave refused consent, a wrong purchase goal and a substituted order amount untouched.
These tools work on one layer of measurement: the route data takes from a visitor’s browser to Google. Our guide to attribution for online stores explains why Google Ads, GA4 and your store still report different numbers for the same sale. Here we compare the gateway, server-side tagging and enhanced conversions, then look at what sits outside their reach.
What does Google tag gateway change?
Google tag gateway changes where the Google tag loads from. In a standard setup, your page requests the tag from a Google domain, and the tag sends its measurement requests, or hits, straight to Google. With the gateway, your site loads the tag from your own domain, and some of those hits reach Google through that domain. That is how Google describes it in Set up Google tag gateway for advertisers.
The gateway runs on infrastructure you already have: a content delivery network (CDN), a load balancer or a web server. Google’s developer documentation for the gateway gives no figure for how much data it recovers. Google’s guide to pairing the gateway with a tagging server describes the benefit in general terms. Serving scripts from your domain and sending data to it, the guide says, makes the setup more durable and the data more complete.
On Cloudflare, the gateway is free on any plan, according to Cloudflare’s launch announcement from May 2025. The tag already on your page stays as it is. The step-by-step setup is in our Google tag gateway setup on Cloudflare.
Cloudflare’s announcement also reports that early testers saw an average 11% uplift in data signals. That is a vendor’s beta average with no published method, and it counts data signals, not orders.
What stays the same matters just as much. The page still runs the same Google tag, so it collects the same data under the same consent rules.
The script in the browser still creates cookies, now under your domain (Cloudflare’s announcement). In Safari, WebKit deletes all cookies created in JavaScript after 7 days without user interaction with the site (Tracking Prevention in WebKit). The gateway leaves that limit in place.
What does server-side tagging add?
Server-side tagging moves tag processing from the browser to a server. Instead of the page sending hits to each platform, the browser sends them to a tag container running on a server you control. According to An introduction to server-side tagging, only you have access to the data on that server until you send it on. You decide what shape it takes and where it goes.
Control is what a server adds on top of the gateway. Google’s Google tag gateway guide spells it out: you can cleanse, enrich and control what the container forwards to Google and to other ad or analytics platforms. In Google’s server-side guide to Google Ads conversions, the server tag can take the conversion value from your own database instead of the page. Google’s example is a variable named Profit lookup.
Cookies can change too. Google’s Custom domain configuration page says a tagging server on your own domain or subdomain gets the security and durability benefits of server-set cookies. On the cloud provider’s default domain, it can only set JavaScript cookies. Google calls the same-origin setup the best practice: the server sits on a path of your own site, such as www.example.com/metrics.
Safari adds a condition. WebKit caps server-set cookies at 7 days when it detects third-party CNAME or IP-address cloaking. One example is a subdomain whose DNS points to another company’s host.
The price is a server and someone to run it. Google’s guide Set up server-side tagging with Cloud Run puts each server at approximately $45 a month. It recommends at least 2 servers to reduce the risk of data loss. That is about $90 a month before any setup work.
The gateway and a server work together. On its Google tag gateway for advertisers page, Google writes: “We recommend you complete both steps for the most durable tagging setup.” In the combined setup, your CDN serves Google’s scripts from your domain, and the server processes the data before it goes anywhere.
Where do enhanced conversions fit?
Enhanced conversions solve a different problem: matching. They add hashed customer data, such as an email address or phone number, to the conversion. That gives Google another way to connect an order to an ad click besides cookies.
The Google Ads API page Manage online click conversions requires SHA-256 hashing for this data, which turns each value into a one-way code. The page names three ways to set up enhanced conversions: Google Tag Manager, the Google tag or the API.
With the API route, you can send the customer data within 24 hours of the conversion, so it can come from your CRM or customer database. Whichever route you pick, enhanced conversions work alongside the gateway or a server. Setup steps are in our enhanced conversions setup guide.
Enhanced conversions also depend on consent. Google’s Consent mode overview lists their hashed first-party data among what switches off when ad_user_data is denied. So if few of your visitors give consent, you get less from enhanced conversions, whatever the route. If you have switched them on and the numbers stay flat, start with enhanced conversions diagnostics.
Gateway, server-side tagging and enhanced conversions side by side
Each of the three works at a different point, so you can run all three at once.
| Google tag gateway | Server-side tagging | Enhanced conversions | |
|---|---|---|---|
| What it changes | Where the tag loads from and the path of some hits | Where hits are processed before any platform gets them | How Google matches a conversion to an ad click |
| Where it runs | Your CDN, load balancer or web server | A tagging server you run, for example on Cloud Run | The Google tag, Google Tag Manager or the Google Ads API |
| Can it change the data | No: same tag, same data | Yes: clean, enrich, choose destinations | Adds hashed customer data to the conversion |
| Cookies | Set by the script, under your domain | Server-set cookies when the server sits on your domain | No change |
| Platforms | Google tags | Google and other ad or analytics platforms | Google Ads |
| Running cost | Free on Cloudflare | About $45 per server a month, at least 2 recommended | No server of its own |
Consent decides more than the route
Consent mode sets what Google tags may do before and after a visitor chooses. Google’s consent mode overview describes two modes. In basic mode, tags stay blocked until the visitor interacts with the banner; if they refuse, nothing reaches Google, not even the consent status. In advanced mode, tags load with consent denied by default and send cookieless pings until the visitor agrees.
Those pings feed modeled conversions: Google estimates the conversions it could not observe. Basic mode gets a general model, advanced mode an advertiser-specific one. Google models conversions only when your tag meets a data collection threshold.
The gateway serves the same Google tag, which Google lists among the tags with built-in consent checks. A visitor’s refusal holds, whichever domain the hit travels through. A server can only process what the browser sent it. Our guide to Consent Mode v2 for online stores shows how to test your banner and both modes.
The conversions you recover are sales you already had
The gateway, server-side tagging and enhanced conversions change counting and leave demand as it is. The extra conversions after a switch are orders that already happened and now show up in the report. Your store’s order list stays exactly the same.
Example store, not client data.
A tableware store’s Google Ads report shows 300 orders and 180,000 in revenue on 40,000 of spend, a ROAS of 4.5. After the gateway goes live, Google Ads reports 5% more of the same orders: 315 orders and 189,000 in revenue. Reported ROAS rises to about 4.7. The store’s own system shows the same sales as before.
Two things follow. First, treat any jump across the switch as a change in measurement rather than performance, and check reported orders against your store’s orders. Second, a target ROAS from before the switch assumed fewer counted orders, so the same target is now easier to hit on paper. The gain you pay for is indirect: bidding learns on a fuller set of your real orders.
What none of them fixes: the goal and the order amount
Infrastructure decides whether a conversion reaches Google. Whether that conversion is true is a separate question. Four errors pass through the gateway, a server and enhanced conversions untouched:
- The primary goal is a funnel step. It is an add to cart, a checkout start or a page view, so Google optimises started actions instead of paid orders.
- Every order carries the same amount. The conversion settings hold one fixed number, and that number goes to Google instead of the order amount. Revenue in the report becomes the number of orders multiplied by it.
- One order is counted twice. A second primary goal, such as an import from analytics, an old tag or a goal from your shop platform, fires on the same order.
- One payment is many times larger than a normal order. A wholesale shipment, a test order or a checkout glitch looks to Google like a success worth repeating.
Server-side tagging is the only one of the three that can change the conversion value, and only if someone builds that step. By default, Google’s server-side Google Ads tag reads the transaction ID, conversion value and currency from the e-commerce data it receives. A wrong amount on the page reaches the server as the same wrong amount.
These errors live in the Google Ads account, so that is where to check them. The portal’s conversion check reads your account and asks four questions:
- Do your ads learn on a purchase?
- Does the real order amount arrive with each order?
- Do two goals count the same order?
- Are some orders many times larger than your average order value?
The portal treats a product’s conversion value for a day as an anomaly when it is more than 30 times the account’s median order value. It also flags a product campaign with at least 10 conversions if their average value falls below 0.3× or above 8× that median.
The portal reads your Google Ads account and shows what your ads are learning on. To see whether the gateway works, check the site itself in the browser’s network tab.
In our check of 8 store websites, no fault was about the route
From March to June 2026, we ran an automated check on the websites of 8 online stores: seven home pages and one cart page. The script opens the page, tries to accept the cookie banner and records the tracking requests the page sends. We checked some sites more than once; a site counts if the issue showed in any of its checks.
| What the check found | Sites | Does a gateway change it? |
|---|---|---|
| Page views sent to two or three different GA4 measurement IDs | 4 of 8 | No, it passes each one through |
| The tag container loaded, but the check recorded no GA4 or Google Ads hit to Google’s domains during the visit | 4 of 8 | No, it can’t make a tag fire |
| Of those 4: no such hit even after the test accepted cookies | 2 | No |
Several GA4 IDs can be deliberate, for example a roll-up property that collects data from a group of stores. They can also be a leftover from an old setup, and only the owner can tell which.
One site shows both findings. In April it sent hits to three GA4 IDs. By the late-May check a consent layer had appeared, and that check and one in June recorded no GA4 or Google Ads hit.
The check has limits. The script recognises hits sent to Google’s own domains, so a hit routed through the store’s domain, for example to a tagging server, would not show up. Each check is one automated visit, so it shows what a page sent, not what the store’s reports contain. This is an observation on 8 sites, not a market sample.
See what share of your numbers you can trust. Sign in with Google in one click. The portal changes nothing without your consent.
Which one is worth it at your size?
Start with the gateway. If your domain is already on Cloudflare, it costs nothing and the tag stays as it is. Other CDNs and load balancers can run it too, with more setup work. Google’s setup guide recommends the gateway for the most durable tag configuration.
Server-side tagging pays off when you need something only a server does and you have someone to maintain it:
- Your own conversion value. For example, margin from a database instead of the amount on the page.
- One data stream for several platforms. Google and other ad or analytics tools fed from the same server.
- Control before data leaves. You choose what each platform receives.
Hosting is a fixed cost, so how much it matters depends on your size. Of 114 stores in our data, 84 had median monthly Google Ads revenue under $10,000 (GetProfit data, June 2025 – June 2026). Here is what the two servers Google recommends cost as a share of that revenue:
| Monthly revenue reported by Google Ads | Stores in our data | ~$90 of hosting as a share of that revenue |
|---|---|---|
| Under $2,000 | 22 | More than 4.5% |
| $2,000–10,000 | 62 | 0.9–4.5% |
| $10,000–30,000 | 20 | 0.3–0.9% |
| $30,000 and more | 10 | 0.3% or less |
Setup work and maintenance come on top of hosting. Until one of these three needs applies, start with the cheaper layers: the gateway, enhanced conversions and a consent setup that works.
What to do, in order
- Check the goal and the order amount first. The primary goal should be a purchase, and each order should carry its real amount. Otherwise the gateway, a server and enhanced conversions only deliver wrong numbers more completely.
- Count your GA4 IDs. Open your home page with the browser’s developer tools, filter network requests by
collectand note every GA4 ID (G-…). Each extra one needs a reason. - Test consent. Load the page in a private window without accepting cookies, then accept, and watch what fires. In basic mode nothing should go out before consent; in advanced mode you should see cookieless pings.
- Switch on the gateway. On Cloudflare, follow our setup guide. Then check in the network tab that the tag loads from your domain.
- Switch on enhanced conversions. Make sure your banner passes
ad_user_dataas granted when the visitor agrees. - Mark the date. Add a note in your reports, compare reported orders with your store’s orders, and review targets set before the switch.
- Add a server only for a named need. That need is your own conversion value, several platforms fed from one stream, or control over what leaves your site.
Sources
- Set up Google tag gateway for advertisers — the tag loads from your domain and some measurement requests go through it; runs on a CDN, load balancer or web server; recommended for the most durable tag configuration. Checked 2 October 2026.
- Google tag gateway for advertisers — Google recommends completing both the gateway and server-side steps. Checked 2 October 2026.
- Google tag gateway guide — with the gateway and a tagging server together, serving scripts from and sending data to your domain gives more complete data; a server container can cleanse, enrich and control what is forwarded. Checked 2 October 2026.
- An introduction to server-side tagging — the server container runs on a server you control; only you see the data until you send it on. Checked 2 October 2026.
- Custom domain configuration — server-set cookies on your own domain or subdomain; JavaScript cookies only on the default domain; same origin as best practice. Checked 2 October 2026.
- Set up server-side tagging with Cloud Run — approximately $45 a month per server; at least 2 servers recommended. Checked 2 October 2026.
- Google Ads conversions — the server-side tag reads transaction ID, value and currency from e-commerce data; the value can come from a database. Checked 2 October 2026.
- Manage online click conversions — enhanced conversions for web: SHA-256 hashing, setup through Tag Manager, the Google tag or the API, data within 24 hours. Checked 2 October 2026.
- Consent mode overview — basic and advanced modes, modelling and its threshold, built-in consent checks in the Google tag, enhanced conversions off when
ad_user_datais denied. Checked 2 October 2026. - First-party tags in seconds: Cloudflare integrates Google tag gateway for advertisers — vendor announcement, May 2025: free on any Cloudflare plan, an 11% average uplift in data signals among early testers, cookies set by the script under your domain. Checked 2 October 2026.
- Tracking Prevention in WebKit — 7-day limit on cookies created in JavaScript; 7-day cap on server-set cookies under CNAME or IP-address cloaking. Checked 2 October 2026.
- GetProfit data: 114 stores, June 2025 – June 2026 — stores by monthly revenue reported by Google Ads.
- GetProfit tag check: 8 online stores (seven home pages, one cart page), March – June 2026 — GA4 measurement IDs and pages that sent no hits to Google’s domains.
- GetProfit portal methodology — the conversion checks and anomaly thresholds.
Ecommerce Attribution: Google Ads vs GA4 vs Your Shop
Find out which sales number to use for bidding, channel budgets and profit, and why Google Ads dates a sale by the click and your shop by the order.
Enhanced Conversions for Leads — setting it up through Google Tag Manager
When the deal closes not on the site but in the CRM 2 weeks later, ordinary Enhanced Conversions doesn't help. A step-by-step EC for Leads setup through GTM with offline conversion import.
Google Ads Conversion Tracking an Online Store Can Trust
Get purchase, revenue and ROAS numbers you can rely on: make the paid order your only primary conversion, sent with its real amount and a unique order ID.
Google Tag Gateway via Cloudflare — a detailed guide
How to switch Google Tag Gateway on via Cloudflare in 10–15 minutes and recover 11–25% of lost conversions in Google Ads and GA4. Step by step, in English.
Enhanced Conversions — a step-by-step setup for Google Ads and GA4
Form data is already collected on your site but never reaches Google Ads and GA4? Step by step — where to click, what you'll see, how to verify. Setup through Google Tag Manager.
How to set up conversions in Google Ads with Google Tag Manager
A simple and quick manual for setting up conversions.