How GA4 assigns credit for online store sales, which attribution models and lookback windows you can choose, and why GA4, Shopify and Meta never match.

GA4 Ecommerce Attribution: How It Works and Where It Misleads

GA4 attributes ecommerce revenue by giving credit for each purchase to the channels and clicks on the customer’s path, using the attribution model set on your property. Today that means data-driven attribution by default, with two last-click alternatives. It will not match Shopify, Google Ads or Meta, and it isn’t meant to. This guide explains what GA4 does with your orders, which settings matter, why the numbers disagree, and how to use them to make budget decisions.

Moormarketing

Turn Attribution Into Better Growth Decisions

Moor Marketing develops tailored, data-driven strategies for eCommerce businesses, with hands-on mentoring to help tackle specific growth obstacles.

Explore Moor Marketing

TL;DR

GA4 offers three reporting models: data-driven, paid and organic last click, and Google paid channels last click. First click, linear, time decay and position-based are gone from GA4 reporting.
Changing the model changes reports retrospectively, but only those using event-scoped dimensions such as Source, Medium and Campaign.
GA4, Shopify, Google Ads and Meta will always report different totals. Treat your store’s orders as the truth and use each platform to explain them.
Most “attribution problems” are tracking problems: duplicate purchase tags, missing transaction IDs, missing currency.
Use attribution to compare channels over a consistent window, not to defend a single number.

What GA4 attribution actually measures
The attribution models you can choose in GA4
Where to change your attribution settings
Why GA4, Shopify, Google Ads and Meta never agree
Fix the tracking before you debate the model
Using GA4 attribution to make budget decisions
Our perspective
Get a clear read on your own numbers
FAQ
Sources

What GA4 attribution actually measures

Google defines attribution as assigning credit for an important action, such as a purchase key event, to the ads, clicks and other touchpoints on the customer’s path. The attribution model is the rule set or algorithm that decides how that credit is split.

For a store, the raw material is the ecommerce events your site sends. Google’s developer documentation lists the typical ones: viewing item lists and product details, selecting items, adding or removing products from the cart, beginning checkout, purchasing, refunding and applying promotions. Attribution then takes the purchase event and works backwards through the visitor’s earlier interactions.

Two details trip up store owners. First, Google separates event-scoped and item-scoped metrics: one add_to_cart event can contain several items, so counting events and counting products give different answers. Second, the ecommerce scopes work in layers. Event-level parameters such as transaction_id, value and tax describe the whole transaction, while the items array describes the products inside it. If you mix the layers in one report, expect confusing totals.

Also expect a lag after setup. Google says ecommerce data normally begins appearing within 24 to 48 hours of users visiting the tagged site, so don’t judge a fresh implementation after an afternoon.

The attribution models you can choose in GA4

A lot of older guides still describe six or seven models. Google’s current help lists three models in GA4 Attribution reports, and states that first click, linear, time decay and position-based have not been available since November 2023.

Model	How credit is assigned	Best used for
Data-driven attribution	Machine learning compares converting and non-converting paths to estimate each touchpoint’s contribution	Default view of how channels work together
Paid and organic last click	100% to the last click, paid or organic	A simple, auditable view that is easy to explain
Google paid channels last click	100% to the last Google Ads click, when one exists	Judging Google Ads on its own terms

Data-driven attribution is the default. Per Google, it can weigh factors such as the time to key event, device type, the number and order of ad interactions, and creative type, and it estimates each touchpoint’s contribution by comparing what happens with and without it.

The practical catch: a model that uses machine learning is harder to sanity check. If you want to know why a channel gained or lost credit, a last-click view alongside it is the quickest way to see whether the shift is real or a modelling effect.

Where to change your attribution settings

A property administrator sets the reporting model, the channels eligible for credit and the lookback window under Admin > Data display > Events > Attribution settings, according to Google’s attribution settings page.

Three things to know before you touch them:

It rewrites history in reports. Google says a change to the reporting model applies to historical and future data. The raw collected data is untouched, so you can switch back.
It only affects some dimensions. Reports using event-scoped traffic dimensions (Source, Medium, Campaign, Default channel group) change. User-scoped and session-scoped dimensions such as Session source or First user medium do not.
Lookback windows matter. The default is 30 days for acquisition key events and 90 days for other key events. The alternatives are 7 days for acquisition events, and 30 or 60 days for the others.

For most stores, the window should roughly match how long people take to buy. A $30 consumable might justify a shorter window. A $2,000 sofa deserves the 90-day default, because shoppers return several times before ordering.

Why your Traffic acquisition report doesn’t follow the model

This catches many owners out. The Traffic acquisition report is session-scoped, and Google confirms that sessions use non-direct last-click attribution, with the lookback set by the “All other conversion events” setting (90 days by default). If someone clicks an ad, then returns directly a week later and buys, the earlier non-direct source can be kept for that session, using signals such as gclid, UTM parameters and the referrer.

So if you switch to data-driven attribution and your session-based report doesn’t move, nothing is broken. You are looking at a different scope.

Why GA4, Shopify, Google Ads and Meta never agree

The single most common complaint from store owners is that the platforms add up to more orders than the store took. Search Engine Land lists the reasons: ad platforms can each claim the same conversion, attribution dates differ, cross-device identity resolution differs, and privacy controls or consent banners reduce what can be observed. Its recommendation is sensible: treat the commerce backend as the source of actual orders, and use analytics to explain them.

One mechanism is worth spelling out. An ad platform pixel can claim all the credit for a purchase because it cannot see the other channels that touched the shopper, while GA4 sees a broader journey. Each platform is grading its own homework.

What Shopify does differently

If you run Shopify, there’s a third view. Shopify’s Marketing reports default to last non-direct click, giving 100% of credit to the last non-direct channel. You can also choose last click, first click, any click and linear.

Any click deserves a warning. It gives full credit to every channel that was clicked, so credited sales can exceed real orders. Shopify says it is meant for analysing individual channels or reconciling channel-reported attribution, not for totals. Linear and any click were introduced for Marketing reports in June 2023.

A quick map of the usual gaps:

Source	What it reports	Typical reason it differs
Shopify orders	Actual orders placed	The baseline
Shopify Marketing reports	Orders by channel, in your chosen model	Different model and lookback to GA4
GA4	Key events attributed by the property’s model	Consent, ad blockers, data-driven modelling
Google Ads	Conversions credited to Google ads	Own windows, own model
Meta	Conversions credited to Meta ads	Own windows, claims shared conversions

Fix the tracking before you debate the model

Before blaming attribution, check that purchases are measured cleanly. Search Engine Land warns that duplicate purchase tags, overlapping GTM and GA4 Create Event configurations, and platform integrations can inflate conversions. Its fix is one implementation per event and unique transaction IDs on every ecommerce purchase.

A short audit checklist, based on Google’s implementation guidance:

Send currency at event level whenever you send value or revenue.
Include every ecommerce parameter you have data for, in the documented event and item formats, so reports populate properly.
Give each item an item_id or item_name, and ideally both. Quantity defaults to 1 if omitted, which quietly understates multi-unit orders when you forget it.
Use DebugView while implementing, and run a test order before launch.
Check one purchase fires once. Compare a handful of GA4 transaction IDs against your order list.

For a deeper walk through the item fields, see our guide to 5 item fields to verify in GA4 ecommerce tracking with GTM.

Using GA4 attribution to make budget decisions

The goal isn’t a perfect number. It’s a repeatable way to compare channels so you move money in the right direction. A method that works for most Australian stores:

Pick one reporting model and keep it. Switching models monthly makes trends meaningless.
Compare two views. Look at data-driven and last click side by side. A channel that looks strong in both is a safe bet. A channel that only shines in one is a question to investigate.
Reconcile to orders monthly. Total revenue from your store, then the share each platform claims. If the platforms claim more than 100% combined, that is overlap, not growth.
Judge channels by role. Shopify points out that first click helps gauge top-of-funnel interest, while last non-direct click shows which channel closed the sale. Prospecting channels rarely win a last-click contest and still earn their budget.
Test, don’t assume. When a model says a channel is weak, pause or cut it in a controlled way and watch store revenue, not dashboard credit.

If iOS privacy changes and consent banners are part of your gap, our post-iOS attribution guide for Australian ecommerce covers the practical workarounds. For Google Ads specifically, how to measure Google Ads performance shows how to read the platform and GA4 numbers together.

Our perspective

We see the same pattern with almost every store that comes to us: the owner has three dashboards, three different revenue figures and no confidence in any of them. The instinct is to hunt for the “right” model. In our experience that is the wrong first move.

We start with the orders. If the purchase event fires once, carries a unique transaction ID, the right currency and clean item data, then most of the disagreement shrinks to something explainable. Only then do we choose a model, and we nearly always keep a last-click view next to data-driven so the team can see what the modelling is doing.

The second thing we do is tie attribution to a decision. A report that doesn’t change where next month’s budget goes is decoration. Our strategists, not outsourced analysts, work through your channel mix with you, so the numbers lead to specific moves: scale this campaign, fix this landing page, stop this spend.

The third is humility about the data. Consent, ad blockers and cross-device journeys mean no model sees everything. Treat GA4 as a good witness, not a judge.

— Liza

Get a clear read on your own numbers

If your GA4, Shopify and ad platform figures don’t reconcile and you aren’t sure where to spend next, that is the work we do with eCommerce brands every day. Our 90 day ecommerce digital strategy maps revenue back to GA4 so each channel’s role is clear, and our 12 Week DOUBLE Your Revenue Challenge puts hands-on mentoring behind the changes.

If you’d rather start smaller, our eCommerce strategy workshops are a focused way to audit tracking and agree a plan with a senior strategist. For paid channel choices once your measurement is sound, see our 2026 guide to paid ecommerce advertising channels.

FAQ

What attribution model does GA4 use by default?

GA4 uses data-driven attribution by default. It applies machine learning to converting and non-converting paths to estimate how much each touchpoint contributed. You can switch to paid and organic last click or Google paid channels last click in the property’s attribution settings.

Can I still use first click, linear or time decay in GA4?

No. Google states these models have not been available in GA4 since November 2023, so guides that describe them as selectable are out of date. If you want a first-click or linear view, Shopify’s Marketing reports still offer both for your store’s own data.

What attribution model does Shopify use?

Shopify’s Marketing reports default to last non-direct click, giving all the credit to the last non-direct channel. You can also choose last click, first click, any click or linear. This is why Shopify’s channel figures rarely line up with GA4’s data-driven numbers.

Does changing the GA4 attribution model delete my data?

No. The change applies to historical and future reporting but doesn’t alter the raw data collected, so you can switch back. Only reports built on event-scoped traffic dimensions change. Session and user-scoped dimensions stay the same.

Why does GA4 show fewer or more purchases than my store?

Common causes are consent banners and blockers reducing tracked sessions, duplicate purchase tags inflating counts, and missing unique transaction IDs. Compare a sample of GA4 transaction IDs against your store’s order list to find which one applies.