Data Foundations · GA4 for eCommerce
We implement the complete GA4 ecommerce event set on Shopify, WooCommerce and Magento, remove duplicate purchase sources and reconcile GA4 against your store orders, so the revenue in your reports is revenue you actually booked.
The starting point
Google Analytics 4 doesn't read your order database. It records the events your site sends, and every ecommerce report is built from them. When the purchase event fires without a transaction ID, fires twice, or sends a different value than your store records, everything downstream inherits the error: channel revenue, funnel drop-off and the conversions Google Ads bids on.
Our job is to make that event stream complete, consistent and checked. That's event tracking done as engineering, not a plugin installed once and forgotten.
What we implement
| Event | When it fires | What it powers |
|---|---|---|
| view_item_list / select_item | Collection or search results shown; product clicked | Merchandising and list performance |
| view_item | Product page viewed | Product reports, remarketing audiences, funnel entry |
| add_to_cart / remove_from_cart | Cart changes from any page or drawer | Cart-rate by product and traffic source |
| view_cart / begin_checkout | Cart opened; checkout started | Checkout funnel and abandonment analysis |
| add_shipping_info / add_payment_info | Checkout steps completed | Where checkout loses people, by device and method |
| purchase | Order confirmed | Revenue, key events, Google Ads conversions |
| refund | Order refunded, fully or partly | Net revenue that matches your books more closely |
Every event carries an items array (item_id, item_name, price, quantity, item_category, item_variant). Purchase also carries transaction_id, value, currency, tax, shipping and coupon, and we document which revenue definition value uses.
Platform specifics
Shopify tracks checkout through customer events. The Google & YouTube app can send GA4 ecommerce events, and a custom pixel can load your own tags. Pixels run in a sandbox, which can affect page location and referrer data.
On WooCommerce, purchase usually fires on the order-received page through a plugin or a data layer push. Customers who pay through an off-site gateway and never return to that page don't trigger it, and reloads can trigger it twice.
On Magento and Adobe Commerce, purchase fires on the checkout success page through an extension or a custom data layer. Full-page cache and multi-store setups are the usual suspects when values are stale or in the wrong currency.
GA4 vs. Shopify
Shopify is your system of record. GA4 is an observation tool that misses sessions when visitors decline consent or block scripts. The goal is a stable gap you understand, and most alarming gaps shrink once you compare like for like.
What you get
Every event, parameter and custom dimension written down, with the revenue definition and the owner of each.
Purchase and the few lead or signup events that matter, marked as key events. Not every click.
Payment gateways listed as unwanted referrals, internal traffic filtered and UTM parameters standardized across email, SMS and paid.
GA4 linked to Google Ads and Merchant Center, with one decision on whether Ads counts conversions from GA4 or its own tag.
Event data retention set to the longest standard option, and the BigQuery export enabled where you need raw data.
Test orders across devices and payment methods, checked in DebugView and against your order export.
Where it fits
GA4 is only as reliable as what feeds it. Most GA4 problems are really data layer or tag management problems, and we fix them there. Not sure how far off your numbers are? A Data Health Check answers that first. Once purchases are right, check your channel economics with the ROAS calculator. All of this is part of Data Foundations.
FAQ
It can send GA4 ecommerce events, including purchases, once it's connected to your GA4 property. That's a reasonable setup for many stores. The problems start when a custom GTM pixel or a theme snippet also sends purchases. We keep one source, then check its events against your orders before relying on it.
Almost always because purchase events fire twice. Look for more than one source sending purchases, such as an app pixel and a GTM tag, and for transaction IDs that appear more than once in a GA4 exploration. On WooCommerce and Magento, reloads of the confirmation page are another common cause.
There is no official standard. As a rule of thumb, once you compare like for like (online orders only, same revenue definition, same time zone and date range), a gap within about 10% is normal for browser-based tracking. A larger, explained gap, for example from low consent rates, is fine. An unexplained or growing gap is the real problem.
Either can work. The Google Ads conversion tag with enhanced conversions is usually more direct for bidding, while importing GA4 key events keeps one definition across tools. What matters is choosing one primary purchase conversion, so Smart Bidding doesn't count the same order twice.
Start with the free audit to see your tracking health from the outside, or book a call to walk through your GA4 property with us.