A conversion-rate increase is usually celebrated. But a merchant whose Shopify conversion rate suddenly changes while orders stay steady may be looking at a measurement update, not at newly persuasive product pages. Shopify rolled out a documented session-measurement change on September 21 to 23, 2026 1. Any comparison that spans that rollout needs one extra check before anyone claims a win or a loss.
The short version
- Conversion rate is orders divided by sessions. If Shopify changes how sessions are counted, the rate moves even when customers behave exactly as before.
- The change can push the rate up or down. Filtering bots lowers sessions. Counting sessions that were previously left out raises them.
- Compare orders and net sales across the date, and set a new conversion baseline from the rollout onward.
What Shopify changed
According to Shopify, sessions now follow a visitor's continuing activity instead of ending automatically at midnight UTC. Sessions without a pageview can be included, for example a customer who goes straight to checkout from a cart link. Shopify also filters identified bot sessions out of the relevant session-based reports by default. A session still expires after 30 minutes of inactivity 1.
Three separate changes, then, act on one number:
| Change | Effect on sessions | Effect on conversion rate |
|---|---|---|
| Sessions continue past midnight UTC | Fewer, longer sessions | Can rise |
| Sessions without a pageview can count | More sessions | Can fall |
| Identified bot sessions filtered by default | Fewer sessions | Can rise |
The net result depends on the store. A store with a lot of bot traffic may see sessions fall. A store with many direct-to-checkout visits may see them rise. Neither outcome means customers changed.
A worked example with zero extra orders
Suppose a historical report shows 10,000 sessions and 150 completed-checkout sessions. Conversion rate is 150 ÷ 10,000 = 1.50%.
If a differently defined session count later shows 9,000 sessions for the same 150 completed checkouts, the rate becomes 150 ÷ 9,000 = 1.67%. That is an increase of 0.17 percentage points, or an 11.1% relative increase, with no additional orders.
Now run it the other way. If the new definition counted 11,000 sessions for the same 150 checkouts, the rate would be 1.36%, a fall of 0.14 percentage points or 9.1% relative. Same customers, same orders, opposite headline.
| Sessions | Completed checkouts | Conversion rate | Change vs 1.50% |
|---|---|---|---|
| 9,000 | 150 | 1.67% | +11.1% |
| 10,000 | 150 | 1.50% | Baseline |
| 11,000 | 150 | 1.36% | -9.1% |
These figures illustrate how sensitive a ratio is to its denominator. They are not an estimate of how much the Shopify update affected any merchant.
How to audit your real reports
- Mark September 21 to 23 on your analytics timeline so the break is visible to everyone who reads a chart.
- Compare orders and net sales for periods on each side of the rollout. These are counted from orders, not from sessions, so the change in session rules does not move them. Completed-checkout sessions are session-based, so read them with the same caution as the rate itself.
- Check the bot filter. Confirm whether a "Human" or bot view is available and applied consistently across the periods you compare.
- Look at custom tracking. Headless storefronts and custom analytics may be affected differently from Shopify's own reports.
- Do not merge sources blindly. Combining GA4 events and Shopify sessions in one chart needs a check that both measure the same thing.
Shopify says historical data was not deleted or fully reprocessed, and bot classification is not available the same way for older periods. A simple year-over-year comparison therefore needs a methodology note, not one blended line 1.
Why agencies and in-house teams should care
A CRO report that credits a new product page or a Meta campaign for this specific shift could claim an improvement that never happened. The reverse is just as costly: a brand might cut a profitable campaign because reported sessions rose under the new definition.
The safe habit is the same either way. Annotate the break in the reporting rules, show orders and revenue next to conversion rate, and set a new baseline after the migration. If a funnel analysis is also running on store conversion, traffic with no sales or add-to-cart with no purchase, the same note belongs on each.
A before-and-after worksheet
Put the two periods side by side and let the order-based numbers speak first. This example is hypothetical: three weeks before the rollout and three weeks after, with the same orders and revenue.
| Measure | Before rollout | After rollout | Change |
|---|---|---|---|
| Sessions | 20,000 | 18,400 | -8.0% |
| Orders | 300 | 300 | 0.0% |
| Conversion rate | 1.50% | 1.63% | +8.7% |
| Net sales | $24,000 | $24,000 | 0.0% |
| Average order value | $80 | $80 | 0.0% |
Orders, net sales and average order value did not move. Only the denominator did, so the rate rose 8.7% in relative terms with nothing changed in the store. Showing the table this way answers the stakeholder question "did we improve?" before anyone writes a story about it.
Metrics that stay comparable
Some measures do not divide by Shopify sessions and can be compared across the date without a note.
- Orders and net sales.
- Average order value (net sales divided by orders).
- Refund rate (refunded orders divided by orders).
- Contribution per order.
- Store-level efficiency: net sales divided by total ad spend.
- Platform metrics built from the platform's own counts, such as clicks and spend.
Anything that divides by Shopify sessions, including conversion rate, add-to-cart rate and revenue per session, belongs on the "changed definition" side of the chart.
If you were running a test across the date
An A/B test that started before the rollout and ended after it needs care. Both variants are measured under the same rules at any one time, so comparing variants within the same days remains fair. Comparing the test period with the weeks before the test does not. If the test used sessions as its denominator, document the change and check that traffic was split evenly across variants on both sides of the date.
How to word the methodology note
Add one sentence to any report that spans the date: "Shopify changed how sessions are counted between September 21 and 23, 2026. Conversion rate before and after is not directly comparable. Orders and net sales are unaffected." Then show those two measures next to the rate. It takes thirty seconds and prevents the wrong conclusion from being repeated for months.
What to do this week
- Add a dated note at September 21 to 23 on every dashboard that shows conversion rate or sessions.
- Pull orders and net sales for the three weeks before and the three weeks after, and put them beside the rate.
- Re-run any funnel diagnosis built on session counts, such as add-to-cart rate or checkout rate, using only post-rollout data.
- Check campaign reports where cost per session or revenue per session was a target, and reset the target from post-rollout data.
- Tell anyone who reads your reports, including clients and your finance lead, what changed and why a chart may show a step.
- Start the new baseline. Give it at least a full trading cycle before drawing conclusions from it.
Teams that skip this tend to repeat one of two mistakes. They credit a campaign or a page redesign for a jump that is only a definition change, or they panic over a drop that is the same effect in reverse. Both are avoidable with the table above.
Rebase your funnel
The tool below takes the sessions and orders from before and after the change and shows the rate under each definition. Use it to see whether a movement comes from the numerator (orders) or the denominator (sessions).
Common questions
Did Shopify change how conversion rate is calculated?
Shopify changed how sessions are measured, and conversion rate uses sessions as its denominator. The documented rollout ran September 21 to 23, 2026 1.
Why did my conversion rate go up with no change in sales?
If fewer sessions are now counted, for example because identified bots are filtered, the same orders divide into a smaller number. The rate rises with no extra orders.
Why did my sessions go up?
Sessions without a pageview, such as direct-to-checkout visits, can now be included. Longer continuous sessions can lower the count at the same time, so the net effect varies by store.
Can I compare this September with last September?
Only with a note. Shopify did not reprocess historical data, so the two periods use different definitions. Compare orders and net sales, and treat conversion rate as a new series.
Sources
- Shopify Help, Changes to sessions and conversion rate in Shopify Analytics help.shopify.com
- Shopify Help, Analytics fields reference help.shopify.com
Nenad Vasic
Founder and CEO of Fenix Network Media. His work focuses on the economics behind customer acquisition, conversion, measurement and profitable growth. Calculations in this article are original; third-party figures belong to their publishers and are credited in the sources.
Rebase your Shopify funnel
Use your own numbers, then bring the result to a conversation if you want a second set of eyes on the plan.
