How to choose a Shopify attribution model, step by step

Usually last non-direct click for the weekly report. To choose, compare it with last click and first click in one Shopify marketing report, then check GA4 and Meta against the same rule. If the models barely differ, fix your tagging before you debate models.

Step by step

Use one full month and the same dates in every tool. In Shopify, staff need the Home, Reports and Marketing permissions to see the Growth section. In Meta, you need administrator, advertiser or analyst access to the ad account.

1. Open a report that carries the Attribution menu. From your Shopify admin, go to Analytics > Reports, click the Category filter and choose Marketing. Open Performance by referring channel. The Attribution menu sits in the configuration panel whenever a report pairs a sales metric with a marketing dimension, and it starts on Last click. Menu path: Analytics > Reports > Category: Marketing > Performance by referring channel.
2. Add last non-direct click next to last click. Select both models in the Attribution menu and read the Direct row first. Last click counts direct visits, while last non-direct click hands each of those sales to the channel before it. The drop in Direct is the slice of direct sales that had an earlier tracked click. Menu path: the report's configuration panel > Attribution.
3. Add first click to see who introduces buyers. Select First click as a third model. Shopify says it shows which channels generate top-of-funnel interest. A channel that looks big under first click and small under last non-direct click starts journeys that others finish. Menu path: the same Attribution menu.
4. Keep linear and any click for narrow questions. Linear gives every click in a journey equal credit, direct visits included. Any click gives each clicked channel full credit, so its totals run past your real sales. Shopify suggests it for a single channel, or to reconcile with what that channel reports. Menu path: the same Attribution menu.
5. Check the Growth section's channel report. Go to the Growth page, find Top channel performance and click View channel report. Its Attribution model setting offers last click, first click and last non-direct click. Shopify's attribution data there starts on 1 October 2021. Menu path: Growth > Top channel performance > View channel report > Attribution model.
6. Put GA4 on the same rule. In GA4, click Advertising, then Attribution models under Attribution, and select your purchase key event. Use the drop-downs in the Attribution model (non-direct) columns to set data-driven against paid and organic last click. Google's help says paid and organic last click and last non-direct click are one model under two names. Menu path: Advertising > Attribution > Attribution models.
7. Write down GA4's reporting model and window. In Admin, under Data display, click Events, then Attribution settings. Note the reporting attribution model and the key event lookback window, which defaults to 90 days for purchases. A new reporting model rewrites past reports as well, so record the current one before anyone touches it. Menu path: Admin > Data display > Events > Attribution settings.
8. Read Meta's numbers by attribution setting. In Ads Manager, click the Columns: Performance drop-down, scroll down and select Compare attribution settings. Pick 1-day click, 7-day click and 1-day view. Meta reports each ad set under its own setting, usually 7-day click and 1-day view, so its totals will not match any Shopify model. Menu path: Ads Manager > Columns: Performance > Compare attribution settings.

A worked example

For illustration, say last month's Performance by referring channel reads, under Last click: Direct 40,000, Google 15,000, Email 10,000 and Instagram 5,000. In this worked example, last non-direct click moves 12,000 out of Direct: 6,000 to Email, 4,000 to Google and 2,000 to Instagram. Direct reads 28,000 in this worked example, and the total stays at 70,000.

Email gained the most, so your emails were bringing back people who then typed your address to buy. For illustration, first click then shows Instagram at 9,000 and Email at 5,000. In this worked example, Instagram introduces more buyers than it closes, and Email closes more than it introduces.

So the house rules write themselves. Last non-direct click runs the weekly report. First click gets a monthly look. And Instagram gets a holdout before anyone cuts its budget, because the two views disagree about it.

One store's anonymised export shows a different outcome. On its Channels sheet, Direct holds 57.7% of revenue under last click, first click and touched views alike. If its Shopify data matched, step 2 would barely move Direct, since a direct-only journey keeps its Direct label under last non-direct click.

The same store's Journeys sheet shows where step 3 still helps. Journeys with 4 to 9 touches took 16.9 days to buy on the Journeys sheet, and journeys with 10 or more touches took 16.0 days. Journeys of that length sit inside Shopify's 30-day reset, so first click would still see where they began. For that store, a new model changes little, and tagging checks and a holdout change more.

What should you check when Shopify and GA4 disagree?

• Direct is far bigger in Shopify. Shopify's report starts on Last click, which counts direct visits, while GA4's models leave direct visits out unless a path was direct only. Compare Shopify's last non-direct click with GA4's paid and organic last click instead.
• First click names different channels. Shopify resets a first interaction after 30 days without a purchase, and after each order. GA4's purchase lookback runs 90 days by default, so GA4 can reach further back.
• A buyer switched devices. Shopify says all storefront data has included cross-device reporting since 1 March 2023. GA4 identifies people through the reporting identity you choose, so it may see two visitors where Shopify sees one.
• GA4's numbers moved after the fact. Google's help says conversions can be reattributed for up to 7 days after the conversion. Treat the most recent week as provisional.
• Shopify's sessions jumped in late September 2026. Shopify says its session measurement rollout runs from 21 to 23 September 2026. That changes session counts and conversion rates, so keep comparisons on one side of those dates.
• Meta's total beats every Shopify model. Meta warns that results across ad sets with different attribution models cannot be compared in its Campaign Overview table. Compare settings inside Ads Manager and channels inside Shopify, never one against the other.

What to do this week

1. Write the house rules where the team will see them. Note one line per job: weekly report on last non-direct click, monthly look on first click, bidding on each platform's own setting. Pass: every report you share names its model in its title. Fail: two reports built on different models land in the same meeting.
2. Run steps 1 and 2 on last month and log the drop in Direct. Pass: you know what slice of Direct had an earlier tracked click. Fail: Direct barely moves, so audit your UTM tags before any model debate.
3. Pick one channel for a holdout. Choose the channel whose first click and last non-direct click numbers differ most. Pass: you have dates and a region split for a test of that channel. Fail: budget moves on model numbers alone, which only reshuffles credit.

Check the homework. Your GA4 Attribution paths export already holds the evidence. Causality Engine reads that one file and shows what each channel caused next to what last-click gave it, in 1 to 2 minutes, for €99 once (excluding VAT), refundable within 30 days. Check the homework

Sources, 1 October 2026: Marketing reports (Shopify); Measuring marketing performance (Shopify); Acquisition reports (Shopify); Get started with attribution (Google); Key event attribution models report (Google); Select attribution settings (Google); Scopes of traffic-source dimensions (Google); Compare attribution settings in Meta Ads Manager (Meta); About attribution models and attribution settings (Meta); Results (Meta); Ad Account Insights (Marketing API reference) (Meta)

Related answers

• All Attribution models questions
• Which attribution model should a Shopify store use?
• How to pick the best ecommerce attribution model in GA4
• Attribution window, defined
• The quarterly attribution checkup