Server-Side Tracking
Server-side tracking sends purchase and lead events from a server you control to Google Ads, GA4 and Meta, instead of trusting a script in the shopper's browser. On stores with a third-party checkout, ad blockers or heavy iPhone traffic, it is often the only way the ad platforms see the orders your ads actually produced.
Prices shown are starting points. The final cost depends on the size and scope of your project, and we confirm it in writing before any work begins.
The usual story goes like this. The store is selling, the bank account agrees, and Google Ads reports a fraction of it. Or the opposite: Google Ads reports more conversions than the store has orders. Either way, Smart Bidding is now optimising on numbers that are wrong, and every rupee of spend follows them.
On Indian Shopify stores the most common cause is the checkout. One-click checkouts such as Shiprocket, Fastrr, GoKwik and Shopflo take the shopper off your theme to finish the order, so the thank-you page where your Google and Meta tags were meant to fire never loads. The order happens. The tag never hears about it. Ad blockers, Safari's cookie limits and consent refusals remove more on top of that.
Server-side tracking fixes this at the source. We run a server Google Tag Manager container on Stape, on a subdomain of your own site, and feed it purchases from the store's order data rather than from the browser. We recover the Google click id for each order, so Google Ads can match the sale to the click that produced it, even when the checkout never loaded a single tag.
We also clean up what is already there. Duplicate conversion actions, a second GA4 property nobody remembers creating, apps that inject their own tag containers, auto-applied recommendations quietly rewriting bids: on one store we found a single misconfigured conversion action that had booked over 6 lakh fake conversions. Fixing the collection only matters once the counting is honest.
What we will not do is promise conversions will double. Only the orders that actually came from an ad click can be credited to one, and on most stores that is well under all of them. The goal is a number you can trust: Google's reported conversions landing close to the orders Google really drove, and not above them.
Case study
An Indian D2C eyewear brand on Shopify, Meta and Google Ads
The problem
A growing eyewear brand on Shopify, selling through a one-click checkout app with a large share of cash on delivery, spent daily on Meta and Google Ads. On the day before we started, Shopify recorded 189 orders and Meta received 114 of them from the store. Purchase events carried the buyer's email on 1.3% of orders, so Meta could match very few of them to real people.
Google had the opposite problem. Four purchase counters were all set as main, so the same order could be counted up to four times. The website tag manager held 95 tags, 59 of them paused, and the server events to Meta were still in test mode, so they did not count for ad delivery at all.
What we did
Every order is now sent to Meta from the store the moment Shopify records it, with the buyer's details prepared the way Meta expects, and browser and server share one order id so Meta counts it once. Test mode was switched off and an old domain that was feeding the pixel was blocked.
In Google Ads one purchase counter stays main and the others observe only, with a new order based counter alongside. Automatic bid and target changes the owner had not chosen were switched off. A privacy safeguard strips health related words from what is sent to Meta, using the brand's own word list.
Nothing that fired was paused without the owner's approval. Every change went live as a numbered version with a one click way back, and we compared Shopify orders against Meta every half hour during the rollout.
The result
On the first full day after the fix, Meta received 219 of 219 Shopify orders from the store. The browser alone would have caught 123. Meta's match score for Purchase rose from 6.1 to 8.2 out of 10, with email and phone on every Purchase event.
Google now reports fewer sales than before, which is the correction and not a drop in sales. We did not promise more orders or a better return from ads, and we do not promise it here. What changed is that both platforms now learn from every real order, counted once.
What's included
- Tracking audit: every tag, container, pixel and conversion action on the site and in the ad accounts
- Server GTM container on Stape with a first-party subdomain
- Google Ads conversion tracking from real order data, click id included
- Enhanced conversions and Meta Conversions API with browser and server deduplication
- GA4 purchase events that match the store's orders
- Fix for third-party checkouts (Shiprocket, Fastrr, GoKwik, Shopflo) that hide the thank-you page
- Cleanup of duplicate and fake conversion actions, extra GA4 properties and stray tag containers
- Consent Mode v2 configuration
- Request-level proof that each event reached Google and Meta
- Reconciliation of ad-platform conversions against store orders after go-live
What you actually get
Orders the ad accounts can see
Purchases sent server to server from store data, so a checkout that never loads your tags stops mattering.
Proof, not a green tick
We trace individual orders through the Stape logs to the conversion action they landed in, and show you the trace.
An honest conversion count
Fake and duplicate conversion actions removed, so Smart Bidding learns from sales, not noise.
Everything in your accounts
Your Stape container, your GTM, your Google Ads and Meta. You own every piece and every login.
How it runs
- 01
Audit
Map what fires today, compare ad-platform conversions against real store orders, and find where the gap comes from.
- 02
Clean
Remove duplicate and fake conversion actions, stray containers and conflicting app tags.
- 03
Build
Stape server container, first-party subdomain, order-based purchase events for Google Ads, GA4 and Meta.
- 04
Prove
Trace real orders end to end and reconcile a clean measurement window against store data.
Questions people actually ask
Google told us conversion tracking is impossible with our third-party checkout. Is it?
Not with server-side tracking. The browser tag cannot fire on a checkout page you do not control, but the order still lands in your store, and the click id can usually be recovered from it. We send that order to Google Ads from the server, so the checkout provider no longer decides what Google sees.
Does this work on Shopify without Shopify Plus?
Yes. We do not need checkout extensibility or Plus-only features, because purchase events are built from order data rather than from scripts on the checkout page.
Why Stape and not Google Cloud?
Stape hosts server Tag Manager containers for a flat monthly fee, adds a first-party custom loader and Cookie Keeper, and keeps the logs readable when something needs proving. Google Cloud works too, but it costs more to run and more to maintain for the same result on most Indian stores.
Google Ads shows more conversions than we have orders. Can you fix that?
Yes, and it is more common than under-counting. It usually means a duplicate conversion action, a page-view counted as a purchase, or a codeless action counting button clicks. We find which action is inflating the number and remove it before building anything new.
How much will conversions go up?
We do not promise a number. Only orders that came from an ad click can be credited to one. What we promise is that the reported number will match reality, and we show you the reconciliation against your store's orders.
What does Stape hosting cost?
Stape bills you directly on its own monthly plans, which scale with request volume. We recommend the smallest plan that fits your traffic and set it up in your name.
Is server-side tracking a way around privacy rules?
No. Consent still applies wherever the tracking runs. We configure Consent Mode so server-side events respect what the shopper agreed to.

