Table of contents
Google Analytics 4 tells the truth about your store only when WooCommerce sends it correct e-commerce events: view_item, add_to_cart, begin_checkout, add_payment_info, purchase — with value, currency, and item list. Based on these, eight reports answer the questions that determine your margin: where customers drop out of the funnel, which products sell from which sources, how much an order costs from each channel, and whether Black Friday paid off. Without correct events, GA4 shows traffic — not sales.
Thesis: 7 out of 10 stores we audit have GA4 with incorrect purchase tracking
Symptoms: GA4 revenue differs from the store dashboard by 30–50%, purchase fires twice (thank-you page refresh), cart events are missing, and “(direct)” shows up as the main source of orders because the payment gateway “stole” the attribution. The owner makes advertising decisions based on data that doesn't reflect reality.
Configuration matters more than reports. That's why we start with it.
Configuration — 6 checkpoints
| Checkpoint | What to check | How |
|---|---|---|
| E-commerce events | view_item, add_to_cart, remove_from_cart, begin_checkout, add_shipping_info, add_payment_info, purchase | GA4 → Reports → Monetization → E-commerce purchases; DebugView with a test order |
Parameters purchase | transaction_id, value, currency, items with price and quantity | DebugView |
| Deduplication | purchase once per order, even after a refresh | Test order → refresh the thank-you page → check that the event wasn't duplicated |
| Gateway attribution | Returning from the gateway doesn't create a new "referral" session | GA4 → Admin → Data Streams → Exclude referral domains: payment gateway domains |
| Consent Mode v2 | Consent signals passed; modeling enabled | Tag Manager / CMP plugin; context in our article on the cookie banner |
| Internal traffic filters | Your IP and office IP excluded | Admin → Streams → Define internal traffic |
Events are sent from WooCommerce by a GA4/GTM plugin for WooCommerce or by a data layer in Tag Manager. Selection criteria: support for all events with items parameters, and deduplication purchase.
8 reports — every week, 30 minutes
1. Purchase funnel (Explore → Path / Funnel)
Steps: view_item → add_to_cart → begin_checkout → add_payment_info → purchase. How to read it: where the biggest drop-off is. Typically, 60–70% drop off between the cart and checkout (shipping costs, account registration) and 20–30% during checkout. What to do about it — see our article on cart abandonment.
2. E-commerce purchases (Monetization)
Products: views, add-to-carts, purchases, revenue. How to read it: products with high view_item and low add_to_cart point to a product page problem (price, photos, description, availability); details in our article on the product page.
3. Traffic acquisition with revenue (Acquisition → Traffic acquisition)
Channels (organic, paid, social, email, direct, referral) × sessions × purchase × revenue. How to read it: conversion per channel. A channel with lots of traffic and zero conversions is either attracting the wrong people — or being tracked incorrectly.
4. Order acquisition cost per channel
GA4 + Google Ads costs (automatic import) and costs from other sources entered manually. Formula: channel cost ÷ number of orders from the channel. Compare it with the margin on your average order — it's the same calculation we described for Ceneo and Google Shopping.
5. Landing pages (Engagement → Landing page)
Which landing pages lead to purchases and which lead to exits. How to read it: categories and blog articles with organic traffic but no click-throughs to products — internal links to your offer are missing.
6. Store search (view_search_results)
Search queries, number of results, conversion after search. Queries with zero results give you a list for synonyms and product range decisions; we covered this in our article on empty states.
7. Devices and performance (Tech → Devices)
Mobile vs desktop conversion. If mobile converts half as well — it's a checkout or mobile performance problem, not "those kinds of customers."
8. Returning customers (Retention)
Share of revenue from returning customers and time to second purchase. Reading: if returning customers account for <15% of revenue — the store pays for every customer from scratch; room for a newsletter, loyalty program, remarketing.
Comparisons worth setting up once
- Year over year for seasonal peaks (Black Friday 2026 vs 2025) — with the same channels.
- “Buyers” segment vs “non-buyers” — how they differ (source, device, pages).
- Alert:
purchasedrops 30% day over day → email. Catches payment gateway outages and checkout errors faster than your customers do.
What GA4 won't tell you
- Margin — revenue isn't profit; you calculate margin outside GA4 (spreadsheet, system).
- Full attribution after consent is declined — Consent Mode models it, but doesn't reconstruct it.
- Why the customer left — that's what heatmaps and session recordings are for.
- Whether a change helped without comparing periods or running a test — more on that in the article on A/B testing.
Most common mistakes
purchasedoubled by a refresh of the thank-you page.- Payment gateway as an order source — referral exclusion is missing.
- Missing
itemsin events — the product report is empty. - Internal traffic (the owner testing the store) in the data.
- GA4 installed twice (plugin + GTM) — everything is doubled.
- No Consent Mode — after consent is denied, you get zero data instead of modeling.
From our experience
With every store launch, we place a test order with DebugView open — and check purchase after refreshing the thank-you page. That's 10 minutes that saves months of bad data.
We set up GA4 with full e-commerce tracking, Consent Mode and payment gateway exclusions as part of our WooCommerce stores, and we go through the reports with clients during SEO and ad campaigns.
FAQ — GA4 for Your Store
Most often: the purchase event doubled by a refresh of the thank-you page, some users not giving consent (without Consent Mode, their orders aren't recorded), canceled or returned orders (GA4 doesn't subtract them) and internal traffic. A difference of up to 10–15% is normal; anything bigger means you should check your configuration.
Without Consent Mode, no. With Consent Mode v2, GA4 collects data without cookies and models conversions for users who declined consent. It requires a banner that passes the consent state to Google.
In GA4, turn on DebugView (Admin → DebugView) and place a test order. You should see, in order, view_item, add_to_cart, begin_checkout, add_payment_info and purchase with the parameters value, currency, transaction_id and a list of items.
Author: Adrian Kamiński — Websky Studio, Poznań. We build WooCommerce stores with GA4 configured starting from events, not reports — and we check purchase after a refresh.
Your feedback
Was this article helpful?
One click — it helps me choose future topics.

