Shopify's attribution models, explained

By the HonestTag team · Published October 31, 2026 · Updated November 11, 2026

Same orders, same week, a different model setting, and the credit moves.

The five models, in Shopify's words

Model	Who gets the credit (Shopify's rule)
Last non-direct click (the default)	"Gives 100% of attribution credit to the last channel that the customer interacts with before making a purchase, excluding direct visits."
Last click	"Gives 100% of attribution credit to the last channel that the customer interacts with, including direct visits."
First click	"Gives 100% of attribution credit to the first channel that the customer interacts with, including direct visits."
Any click	"Gives 100% of attribution credit to each channel that the customers clicked, recognizing all interactions that contributed to a conversion."
Linear	"Gives equal attribution credit to each click in the customer's journey that contributed to a sale, including direct visits."

Source: Shopify Help Center. Two details in that table catch merchants out. Under the default, "If a customer's journey contains only direct traffic, then their order is attributed as direct, as there's no non-direct source to attribute to." And any click "allocates more credit than orders you've received," which is why Shopify says it is "best used to analyze a single marketing channel."

One order, five answers

A worked example, with made-up numbers. A customer clicks your Meta ad on Monday, clicks a Google ad on Thursday, and on Saturday types your store's address into the browser and buys a $120 order. That is three touches: Meta, Google and a direct visit.

Model	Where the $120 goes
Last non-direct click	Google, $120. The Saturday visit is direct, so it is skipped.
Last click	Direct, $120. The visit nearest the purchase wins, direct or not.
First click	Meta, $120.
Linear	$40 each to Meta, Google and direct.
Any click	Meta and Google each show the full $120, so the channels together claim more than the one $120 order.

Now scale it up. Say 100 orders in a month follow that same path. Under the default, Google shows 100 orders and Meta shows none. Switch to first click and Meta shows 100 and Google none. Nothing about the ads, the spend or the orders changed. Only the rule did.

Why "sales attributed to marketing" is smaller than your sales

The sales attributed to marketing figure on the Growth page has its model fixed. Shopify says sales attributed to marketing "are attributed by using the Last non-direct click attribution model, so direct traffic isn't included" (Shopify Help Center). An order from a customer who only ever arrived direct stays out of it, so the number is expected to be lower than your total store sales. Since October 18, 2026, visits referred by your own store are listed as direct as well, as why Shopify lists your own store as the referrer explains.

Shopify also notes that "Traffic that comes to your store with utm_campaign parameters in the URL is attributed to marketing." On the Attribution page, marketing traffic is the default view, and Shopify says it "includes paid, organic, and unknown traffic."

Why Shopify and your ad platforms disagree

Shopify says its results "might not match the reports in the platforms where you run your marketing, such as Meta Ads Manager or Google Ads" and gives the reason: "Each platform credits sales by using its own rules. For example, if a customer clicks both your email and your Google ad before they buy, then each platform can count the same sale." It also notes that ad spend "can take up to 24 hours to sync to Shopify, so recent ad spend in Shopify can be lower than in the ad platform" (Shopify Help Center).

In the example above, Meta and Google can each report the $120 sale as theirs, while Shopify's default puts it under Google alone. Neither number is broken. They answer different questions with different rules. Ad platform purchases vs Shopify orders walks through the other reasons the counts drift apart. Email works the same way: why Klaviyo's attributed revenue doesn't match Shopify covers opens, windows and send dates.

What a model cannot fix

• Visits that were never recorded with their source. A model only reshuffles credit among the touches the store saw. If a visit was never recorded, or arrived without anything that named its source, every model reads the same gap, so the five can agree with each other and still all miss the same channel.
• New or returning. Credit for an order is not the same as credit for finding a customer. A loyal customer who clicks a retargeting ad before a repeat order gives that ad the sale under the last-touch models. If your question is which ads bring in new buyers, see new vs returning purchase signals.
• Whether the ad caused the sale. Every model hands out credit for orders that happened. None of them tells you whether the order would have happened without the ad.

If all you need is a channel view inside your admin, Shopify's report is already there and costs nothing. Use it with one model you have chosen and written down.

What this means

Shopify says a fair comparison of two reports uses "the same date range, attribution model, and traffic type in both." If those match and the numbers still differ, the gap is real. A channel that holds up under both first click and the default is doing more than one job; one that only shines under last click and last non-direct click is mostly closing sales something else started. Totals above your sales under any click are one order counted on every channel the customer clicked. Shopify's paths reach back up to 30 days before an order, or only as far as the customer's previous order if that was more recent, so a product people weigh for six weeks reads shorter than it is.

Switching models between reports makes a budget call read a change in the rule as a change in performance. Under last click and last non-direct click, a repeat buyer's order goes to the ad that closed it.

HonestTag holds one rule for new customers fixed, counted from your store's own orders with no model switch, so this month and last month are graded the same way.

How HonestTag helps

• One model, stated and fixed. HonestTag credits each new customer to their first recorded ad click or ad-tagged link, counted from your store's own orders. There is no model switch, so this month's numbers are graded by the same rule as last month's.
• On the free Mirror plan, The Gap shows what Google Ads and Meta claim beside the new customers whose first recorded ad click was theirs, over the last 14 days. The Mirror plan sends no purchases to any ad platform. On paid plans, The Gap adds each channel's spend and cost per first-click new customer.
• On higher plans, a read-only table shows what last-click and any-click would have said, per platform, beside first click. You see how each platform's count moves under another rule, without a dial that changes the answer. Those columns are counted forward from the first full day the comparison can count, never rebuilt backward.
• On paid plans, each order HonestTag classifies carries proof: the first recorded ad click and how the order was matched to it, so a number you doubt can be opened and checked.

Why one model and not five is on attribution models, and the rule itself is on first-click attribution.

Install HonestTag free from the Shopify App Store and see, on the Mirror plan, what Google Ads and Meta claim beside the new customers each first brought in.

Frequently asked questions

Sources

• Shopify Help Center: Measuring marketing performance, read October 31, 2026
• HonestTag pricing and plan limits
• HonestTag: attribution models, compared