We rebuilt the funnel's endpoint — orders by landing page — from Shopify's own order data and laid it next to GA4. The short answer: yes, GA4's web-conversion tracking holds up. With one attribution gap, and one big caveat about what GA4 can't see.
Across the top 10 landing pages, GA4 transactions (2,506) and Shopify orders attributed to the same pages (2,279) agree within ~10% over 7 days. The funnel reports are trustworthy in shape.
For pages where people land directly on a product page, GA4 credits 40–60% more orders than Shopify does. GA4 over-attributes PDP-lander conversions — read those with caution.
84% of all Shopify orders carry no web landing page — subscription renewals, returning-direct, app. GA4 only ever sees the ~16% that convert in a web session. Never read total business conversion from GA4.
By landing page, last 7 days (Jun 6–12). Both count only in-session web conversions, so they're directly comparable. Content pages line up closely; the divergence is concentrated on direct-PDP landing pages.
| Landing page | GA4 sess | GA4 tx | Shopify ord | GA4 CVR | Shopify CVR | Agreement |
|---|
GA4 CVR = GA4 transactions ÷ GA4 sessions. Shopify CVR = Shopify orders ÷ GA4 sessions (order truth over web traffic). "Agreement" = how close GA4 tx is to Shopify orders.
The two systems agree on the top content pages within ~10–20% (Rise-2 −11%, Home +7%, Listicle −21%), and across all ten pages they net to +10%. That's well inside normal cross-tool attribution variance — so the LP→PDP and PDP→ATC funnel reports built on GA4 are reliable in shape and direction.
It also re-confirms the Listicle story. The Listicle's order-truth conversion rate (1.47%) is basically identical to Rise-2's (1.40%). The Listicle isn't worse at closing — its weakness is upstream (getting landers to a PDP, the 19% LP→PDP). Two independent data sources now say the same thing.
On pages where traffic lands straight on a product page, GA4 credits far more orders than Shopify does (30-Serving Tin +43%, Coffee Kit +60%, Strawberry Matcha +45%). This is an attribution-definition gap — GA4 counts any session that landed on the PDP and later purchased, while Shopify's order landing-page resolves more conservatively. Net: treat GA4's direct-PDP conversion rates as optimistic. (Rise-2 Short is excluded as noise — 48 vs 5 orders.)
Only 16% of Shopify orders have a web landing page; the other 84% are subscription renewals (Recharge), returning-customer direct visits, and app orders. GA4 tracks ~458 web transactions/day against ~2,275 total Shopify orders/day. This is exactly why blended CAC uses Daily-Spend-Tracker Shopify new-customers, not GA4 — and this report independently validates that choice. GA4 is a web-acquisition lens, not a business-conversion lens.
Sources: GA4 (property 310751693) — sessions & transactions by landing page, Jun 6–12. Shopify via Supermetrics (SHP connector) — order_landing_page_url + order count, same window; full URLs normalized to paths. All processed days.
Why not by hour? Total Shopify orders include subscription/returning volume with its own time pattern, so a raw hourly Shopify-vs-GA4 chart isn't apples-to-apples. The landing-page cut is the clean comparison because orders carrying a landing page are necessarily in-session web conversions. A clean hourly web-orders cut needs a direct Shopify connection (see note).
Note on the data pipe: Shopify's session funnel (LP→PDP→ATC) lives only in ShopifyQL, which has been erroring all week through the Intelligems bridge. This report used Supermetrics' Shopify order data as the workaround. A dedicated Shopify Admin MCP would restore the full funnel directly — recommended next step.
Related: LP→PDP & PDP→ATC by Hour · Funnel Deep Dive.