Rebuilding Google Ads Tracking Behind a Fastrr Checkout

The problem

A womenswear label selling on Shopify had moved to a one-click checkout app. Sales carried on, but Google Ads reported far fewer purchases than Shopify, and the owner had been told the gap could not be fixed.

In one week, 288 of 307 orders went through the checkout app. The customer paid on the app's own page and never reached Shopify's thank-you page, so the browser purchase tag had nothing to fire on.

What we found

Orders created by the app arrived in Shopify with an empty customer journey: no first visit, no landing page, zero sessions. Shopify's own reports agreed with the wrong number, so the gap looked normal. The Google click id had not been lost, but nothing standard was looking where it survived.

The account had bigger problems than the gap. One conversion action had booked 674,022 conversions worth ₹6.42 crore on a store taking a few dozen orders a day, and bidding was learning from it. 21 auto-applied recommendations were switched on. Seven Google tag containers loaded on every page view. The theme head held a broken purchase tag that was one syntax fix away from counting every page view as a sale.

We also checked the denominator before promising anything. Only 49% of orders carried a Google click, and Google's coverage of those was already close to complete. So we told the owner up front that conversions would not double. The job was a number he could trust.

What we built

Every paid order now has its Google click id recovered and is sent to a Stape server-side Google Tag Manager container on the store's own subdomain. The container fires the Google Ads conversion with the order value, the order id for deduplication, and the items bought.

The first version fired, returned success, and still never registered a single conversion. Nothing errored; it looked like reporting lag. We traced it to one thing the server-side tag expects that a server does not send by default, fixed it on every event, and proved the fix at request level on a traced order before calling it done.

We tried email-matched enhanced conversions and stopped within a day: conversions ran at 1.22x the orders that had actually clicked a Google ad, because hashed email credits Meta and organic sales to paid search. The sync now sends only orders carrying a verified Google click id.

Alongside it: the two bad conversion actions removed, auto-apply cut from 21 to 1, duplicate server Meta tags paused because Shopify's own Conversions API already sends those purchases, and the broken theme tag taken out.

The result

Over the first three clean days, Google recorded 105% of the value of the orders that came from Google clicks, inside the ±20% bar we set before starting. The checkout app no longer matters to tracking, and bidding now learns from real sales instead of page events.