Google Analytics 4
Sends GA4 purchase and refund events server-side, from the order rather than the browser.
Overview
Browser-side analytics has two blind spots. It misses the purchases where the tag never fired, and it almost never hears about the returns.
The first makes revenue look smaller than it is. The second makes it look bigger — because a refund that is never reported stays in the numbers as a sale. Together they produce a picture that is wrong in both directions, which is a difficult basis for deciding where to spend.
This app reports from the order. A confirmed order sends a purchase; a return sends a refund as units come back, from wherever the return originated — an agent, an ERP integration, or a resolved claim. Events carry the order number, so a browser tag covering the same purchase collapses into one transaction instead of doubling it.
For brands running several storefronts, one order can report to a group-level property and its own market property at the same time, which is usually what the reporting structure already assumes.
What it does
-
Purchase events from the order
Sends a purchase event when an order is confirmed, so revenue is recorded even when the browser tag is blocked or the shopper closes the tab.
-
Refund events as units return
Returns are reported as refunds from any source — portal, ERP integration or a resolved claim — so reporting reflects what was kept.
-
Multiple properties per order
An order is reported to every enabled property configured for its channel, typically a group-level property alongside the channel's own.
-
De-duplicated against the browser tag
Events carry the order number as the transaction id, so a browser tag firing for the same order collapses into one transaction.
-
Per-property refund idempotency
Each return records which properties have already received it, so a retry after a partial failure never double-deducts revenue.
-
Property management in the portal
Measurement ids, API secrets, channel routing and data region are managed per property in the portal.
Apps that usually run together.
Most operations install more than one. These are the apps this one is most often paired with, because they share the same order.