If your Shopify sessions or conversion rate changed around September 21–23, 2026, check the measurement change before judging store performance. Establish a comparable reporting baseline, then reconcile orders, sales and customer counts separately. B2B inquiry businesses should also check qualified inquiries and sales progress. A different percentage alone cannot establish whether the business improved.
By GRC. Official documentation checked September 24, 2026, Asia/Shanghai. All numerical examples below are teaching demonstrations, not customer results or industry benchmarks.
What changed in Shopify session measurement?
Shopify rolled out the update from September 21 to 23, 2026. Continuing activity no longer splits at UTC midnight; some sessions without a pageview now count; identified bot sessions are filtered by default. A session still ends after 30 minutes of inactivity. Together, these changes can increase or decrease reported sessions. Shopify session update.
That gives you three different explanations to test. Combining a visit that continues past midnight can reduce the count. Including a visit that goes directly to checkout from a cart link can increase it. Removing a previously counted bot session can reduce it again. The net direction depends on the store's activity; there is no universal adjustment factor to apply to an old report.
Read the UML model from session to rules and events. A session groups activity events and follows a measurement policy. Event assessments inform the overall human/bot classification. The diagram explains concepts; it does not document Shopify's internal database or an API schema.
Shopify describes bot classification as conservative: relevant human activity can keep a session in the human category. Filtered reports therefore are not proof that every remaining visit is human. Filtering analytics also differs from blocking crawlers from reaching the site. Do not change crawler access merely to make the sessions chart look cleaner. Shopify bot filtering.
Why can conversion rise without another checkout?
For an online store session conversion rate, use sessions that completed checkout divided by sessions, within the same reporting scope. The numerator is a session count, not a substitute for total orders across the business. Shopify uses this formula in its channel reporting too. Online store conversion definition.
Consider an invented sample with 1,200 sessions, of which 36 completed checkout. Suppose the only adjustment is removing 300 bot sessions that did not complete checkout. Keep all other conditions unchanged.
| Teaching scenario | Sessions | Sessions completing checkout | Session conversion rate |
|---|---|---|---|
| Including those 300 bot sessions | 1,200 | 36 | 3% |
| Excluding those 300 bot sessions | 900 | 36 | 4% |
The rate rises by one percentage point, from 36 ÷ 1,200 to 36 ÷ 900. No extra session completed checkout. This deliberately isolates one denominator effect; it does not simulate the full rollout or assume all bots lack checkout activity. Actual stores can experience changes in both the numerator and denominator.
The measurement update itself does not change total orders, sales or customer counts. Those business totals can still rise or fall because of real trading conditions. Shopify's announcement identifies several affected session metrics, including conversion, add-to-cart, reached-checkout, bounce and pageviews per session. Treat a discontinuity as a reason to investigate, not as automatic evidence of growth or failure. Shopify announcement.
How should you compare dates and filters?
Historical records are not deleted or reprocessed by this update, although filtering can change their presentation. Matching the bot filter does not erase the measurement difference between periods before and after the rollout. Use post-update periods for the new session baseline. Bot classification is available only from October 7, 2025; older sessions cannot be filtered on the same basis. Historical comparison guidance.
Start with one sessions report in Analytics > Reports. Record its name, date range, reporting time zone, device or channel restrictions, and the Human or bot session filter where available. Save those settings beside the export. A screenshot of a percentage without its report context is a poor comparison record.
| Review item | What to record | Why it matters |
|---|---|---|
| Measurement period | Before, across or after the rollout; evidence for the store's status | A mixed window can contain different measurement rules |
| Reporting window | Start/end dates, time zone and full or partial days | Partial days and different weekday mixes distort comparisons |
| Filter settings | Human, Bot, both, or unavailable on this surface | Different selections produce different populations |
| Business context | Promotions, pricing, stock, campaigns and site changes | Real commercial changes may coincide with the update |
| Independent outcomes | Order totals or qualified-inquiry records for the same dates | A denominator shift cannot explain every business result |
Some surfaces, including Home and Live View, do not offer the adjustable filter. Headless implementations may differ. Use a report that supports the control, or record the limitation and have the implementation reviewed if Shopify has flagged a storefront issue. Report controls and implementation notes.
Our suggested review method is to compare two complete post-update windows of equal length and similar weekday structure. For example, September 24–30 and October 1–7 could become candidate windows after confirming the store's update status. They are not automatically comparable: holiday demand, promotions or stock changes still need annotation. Low traffic and long sales cycles require longer observation.
What should B2B and DTC operators reconcile?
Use one reporting discipline but choose outcomes that match the business. A B2B inquiry site may complete its sale through a quotation process outside checkout. A direct-to-consumer (DTC) store can usually inspect a retail order sooner. Counting both through a single website percentage hides that difference.
The review window scopes both datasets. Session metrics retain their measurement version and filters; business outcomes retain their source and definition. Hollow triangle arrows show B2B and DTC outcomes as specializations of the outcome set. This is a proposed review model, not an automatic join between Shopify and sales records.
For a B2B inquiry business, define a qualified inquiry before counting it. A useful definition might require an identifiable company, a relevant need and enough context for follow-up. Deduplicate repeated messages and separate contact clicks from actual inquiries. Then record projects entering quotation and subsequent sales progress, using the team's normal sales-cycle window.
In a second invented example, both periods contain 10 qualified inquiries and four projects entering quotation. If the reported sessions fall, an inquiry-to-session ratio may rise while the project pipeline stays unchanged. That ratio also needs a defensible match between the inquiry population and session source. If the records cannot be connected reliably, show the counts beside each other without claiming precise channel conversion.
For a DTC retail business, reconcile order, sales and customer totals, then inspect payment, cancellation, refund and fulfillment outcomes as needed. If both sessions and retail outcomes move, investigate traffic mix, product availability, pricing and checkout experience. The rollout should not become a blanket explanation for a real trading decline.
Keep GA4 and Shopify exports clearly labeled. Two independent measurement systems are useful comparison evidence, but forcing their session totals to agree is not the purpose of this review. The practical result should be an explainable discrepancy, a consistent baseline and a list of unresolved data gaps.
What belongs in your next reporting note?
Write a short note that another operator can repeat: which report was opened; which dates and filters were used; whether the window crosses the rollout; which independent business outcomes were checked; and which explanation remains unproven. Assign an owner to each gap. Review the setup again after roughly a week, then compare complete windows when sufficient data exists.
If sales also come through AI channels, separate website visits from purchases completed elsewhere. Record the purchase location before using a session-based rate to explain channel results.
GRC provides Shopify design, implementation and operations services. To discuss a reporting or storefront implementation question, bring your store URL, business model, affected report and the date/filter details above. Discuss your project via the WeChat QR or email on our contact page. The next step is a scope discussion; access, deliverables and fees are agreed separately.
Frequently asked questions
Does a higher conversion rate prove the update increased sales?
No. Removing non-converting sessions can raise the rate without adding a completed checkout. Check the numerator, denominator and independent sales records before attributing the change to better performance.
Can I fix an old-versus-new comparison by selecting the same bot filter?
The same filter removes one source of inconsistency, but the underlying measurement method still differs across the rollout. Establish a new baseline from post-update data and label longer historical comparisons with that boundary.
Should a B2B inquiry store use Shopify checkout conversion as its main result?
Only when checkout represents the business outcome being measured. For inquiry-led sales, review qualified inquiries, quotation projects and sales progress. A button click or a higher inquiry-to-session ratio is not evidence of additional qualified projects.
What if the Human or bot session filter is missing?
Check the report and storefront setup. Not every surface exposes the filter. Record what is available, use a supported report for comparisons, and investigate an implementation issue only when there is evidence for one.
Sources and scope
This guide adapts GRC's supplied research and checks the linked Shopify documentation as of September 24, 2026. It does not diagnose a reader's store, validate tracking implementation or predict an uplift. The UML figures are original conceptual models; all calculations and suggested review intervals are illustrative. Recheck current documentation and your own Admin before acting.
