Numbers don't match Shopify
Why Luria's sessions and orders differ from Shopify Analytics or Google Analytics, what should match closely, and when to worry.
Short version: sessions will never match, orders should be close over a long enough window, and a small gap is normal. Here is why.
Why the numbers differ
Different definitions of a session. Shopify and Google Analytics count a session the moment a page loads. Luria counts a session from the first page view where Luria was actually allowed to run. Visitors who block scripts, have Global Privacy Control on, or decline consent are in Shopify's count but not Luria's. Expect Luria sessions to run lower.
Consent and ad blockers. In the EU and UK a meaningful share of visitors decline analytics. Ad blockers remove third-party scripts for another slice. Shopify's own analytics run server-side and are unaffected. Luria is not.
Attributed conversions only. Luria's headline conversion number counts purchases it can tie to a visitor it saw. Orders from a phone call, a draft order, a POS sale, a subscription renewal, or a visitor who never loaded a page with Luria on it are real orders, but not attributed ones. The dashboard labels which is which.
Time zones. Shopify reports in your store's time zone. Luria reports in your dashboard's time zone setting. A day boundary shift moves orders between days without changing the total.
Refunds and cancellations. Luria counts an order when it is paid. Refunds are not subtracted from attributed conversions. Shopify's reports show net figures.
Test orders. Orders placed in Shopify's test mode or with a 100% discount code still count in Luria unless flagged.
Multi-currency. Luria records value in your store's base currency. If you show prices in other currencies, Shopify's report may show converted values while Luria shows the presentment currency converted to base.
What should match closely
| Metric | Expect |
|---|---|
| Total orders over a week or more | Within a few percent of Shopify's paid orders from the online store channel, after excluding POS, draft, and subscription renewal orders. |
| Order values | Same orders, same values, base currency. |
| Attributed orders | Always lower than total orders, by the share of visitors Luria could not see. |
What will not match
- Sessions (see above).
- Conversion rate, because the denominator is different.
- Daily numbers near midnight, because of time zones.
- Anything from GA4, which has its own sampling, thresholds, and modelling on top.
When to worry
Order counts off by more than a few percent over a full week, after excluding non-web orders. Or orders that go to zero while Shopify keeps taking them: that usually means the app's purchase tracking lost its connection.
What to send support
- Your Site ID and store URL.
- The date range, in your store's time zone.
- Shopify's order count for the online store channel over that range (Analytics > Reports > Sales by channel).
- Luria's order count for the same range, with a screenshot.
- Whether anything changed in that window: theme update, app install, checkout customisation, consent banner.