Proof log
Where every numberon this sitecame from.
Every figure traces to a dated screenshot of the client’s own Shopify Analytics. Nothing here is estimated, rounded up or recalled from memory.
01
The figures
Four periods,every stage.
Online-store sessions only. This excludes the store’s other sales channels, so these order counts are smaller than the store’s total orders.
| Period | Sessions | Carts | Checkouts | Orders | Conv. rate |
|---|---|---|---|---|---|
| 1–31 Jul 2026 | 16,349 | 899 | 466 | 234 | 1.43% |
| 1–31 Aug 2026 | 43,520 | 1,076 | 608 | 296 | 0.68% |
| 1–30 Sep 2026 | 18,037 | 1,084 | 719 | 353 | 1.95% |
| 5 Oct–4 Nov 2026 | 15,516 | 1,194 | 1,113 | 615 | 3.96% |
02
Known distortions
Two reasons theconversion rate lies.
Both are visible in the captures above, and both are the reason this site leads with order counts instead.
August’s session count was inflated
August recorded 43,520 sessions against 16,349 in July. The dashboard’s sessions-by-location tile shows Singapore at 21.5K and China at 4.1K for that month; desktop sessions rose 268%. By September, Singapore had fallen to 789 and China to 818.
Conversion rate is orders divided by sessions, so roughly 25,000 sessions that were never going to buy dragged August down to 0.68%. Treat the 0.68% figure as a measurement artefact, not a business result.
Shopify changed how sessions are counted on 21 September 2026
The analytics page carries a notice: “Beginning Sept 21, 2026, sessions and related metrics may shift as Shopify Analytics enables better visibility into real shopper activity.” A Definition change marker appears on the September chart at that date.
Any conversion rate that spans 21 September is not strictly comparable with one from before it. Order, cart and checkout counts are unaffected.
Deliberately absent
No revenue, no customer data, no testimonials I did not earn.
The store’s sales figures exist and I have seen them. They are the client’s to publish, not mine, so no dollar amount appears anywhere on this site.
Nor does anything that could identify a customer. The captures on this page are aggregate analytics screens, which is why they are safe to show you at full size.
Nothing on this site is a projection of what your store would do. It is a record of what one store did, with the parts that are not flattering left in.
Last updated 4 November 2026
Contact
Want this run on your store?
Send the URL and what is not working. I will tell you where I think the orders are leaking before you pay me anything.