Divide your orders by your sessions and multiply by 100. In a made-up example, 180 orders from 12,000 sessions gives 1.5%.
In GA4, this number is called "Session key event rate" and appears in the Traffic acquisition report once it's set to the purchase event. The division is the easy part. The rest is mostly about checking what the tool counts as a purchase.
The benchmark
You get to choose the denominator, and there are three candidates:
| Denominator | The question it answers | Where to find it |
|---|---|---|
| Sessions | Do my channels trigger a purchase during the visit? | Shopify, and the column in the Traffic acquisition report in GA4 |
| Users | What share of my audience bought at least once? | The "Key event rate by user" metric in GA4, almost always higher than the session-based rate |
| People who reached a step | Where does the journey break? | The Purchase journey report or a funnel exploration: purchases divided by checkouts started |
Always say which one you're quoting. The funnel dashboards the authors build count people, Shopify counts sessions, and the difference sometimes has to be clarified mid-meeting.
People love to debate the denominator, but it's the numerator that breaks without warning. The rate GA4 displays counts any session in which a key event occurred, whichever one it is. Purchase is a key event by default.
Google's help page lists seven others, including add_to_cart, begin_checkout and view_item, which are also marked as key events when ad personalization is turned on with a linked Google Ads account.
Going by the authors' reading, your "conversion rate" may therefore count something other than purchases, even if no one on the team marked anything. They haven't measured this on a live property, so check your own with step 3.
The rule
Stick to one denominator from period to period, read it by device and over at least four weeks. Over seven days, the purchase cycle isn't complete and the days of the week don't balance each other out.
Also note the dates when the tool changes how it counts. Shopify changed how it measures sessions between September 21 and 23, 2026, so your rate may have moved without a single order more or less (covered in a separate guide).
The steps
The labels below come from Google Analytics' French help pages, reviewed in early October 2026.
- The column. Open Reports > Acquisition > Traffic acquisition and look for the "Session key event rate" column. If it's missing, the Editor or Administrator role lets you add it. Click the "Customize report" icon (a pencil, top right), then Metrics, Add metric, Apply, and finally Save > Save changes to current report.
- The filter. The column header has an "All events" menu. Choose
purchasethere. - The list. In Admin > Data display > Events, each lit star marks a key event. Note the ones that aren't purchases, so you know what the default setting mixes in.
- The funnel. Reports > Monetization > Purchase journey shows five steps, from
session_starttopurchase, with the drop-off rate under each bar. This report counts people, so its end-to-end rate can't be compared with any rate calculated in sessions. For a step of your own, open Explore > Funnel exploration, which accepts up to ten steps and a breakdown by device category. - The check. Over the same four weeks, divide the purchases GA4 sees by Shopify's orders. The guide on the number to present gives the thresholds for reading the result. Above 100%, a purchase is being sent twice.
The trap
The trap is trusting the rate because the formula is right.
In January 2026, the head of a brand the authors work with was looking at the dashboard they'd been shown, connected to his GA4:
"Today, I figure our conversion rate is around 0.7, 0.8. Here, on the dashboard, I saw it was more like 1.3. [...] So first, let's agree on what our real conversion rate is."
He was right to stop there. Two months later, the cause turned up: two apps were each sending the purchase to GA4, which showed roughly double Shopify's sales. The authors declared their own pre-audit void.
The error also runs the other way. In another case, GA4 saw only six orders out of ten and showed a rate cut in half. On these two sites, the rate varied by roughly a factor of two depending on which tool you opened, which is why step 5's check comes before any reading.
The sentence to present
Copy this sentence, then rewrite it with your own numbers before presenting it to your leadership:
"Our conversion rate is [X]% over the last four weeks: Shopify orders divided by Shopify sessions. GA4 sees [Y]% of those orders. I use it to compare channels with each other, and I report the rate from Shopify."
