7 Shopify Fall 2026 Changes To Apply Before and After BFCM

Shopify published its Fall 2026 Edition on October 28 with more than 150 changes, about a month before Black Friday.

This article sorts the seven that matter by when to act on them. Two need fixing before your code freeze this week. Five can wait until after peak, including two platform deadlines that can break apps in your stack in early 2027.

Your real deadline is the code freeze, not Black Friday

Black Friday is November 27 and Cyber Monday is November 30. Most merchants stop changing their store well before that. We tell clients to freeze between November 10 and November 14, which leaves about a week from now.

A feature you cannot watch perform before peak is a January project. That rules out new ad automation, a checkout layout nobody on your team has seen convert, and any bulk rewrite of product copy.

The change window between today and a typical mid-November code freeze.

What to change before peak is a separate question from how to run the week itself. I covered the margin side with Carla Penn-Kahn of ProfitPeak on The Shopify App Show, including why discount depth should follow when a customer was acquired.

Fix 1: stop comparing conversion rate to last BFCM

This release changed how Shopify counts sessions:

• Bot traffic is filtered more aggressively

• Repeat visits are stitched into unified journeys

• Consent state now affects whether a visit is counted at all

Shopify dates the change to September 21, 2026. It appears as a definition change in your own annotations panel, noting that sessions and related metrics may shift as Shopify Analytics enables better visibility into real shopper activity. Your reported traffic moved without anything changing on your storefront. Sessions and conversion rate are both affected, and the direction depends on how much bot and non-consenting traffic you were counting before.

That matters in three weeks, when somebody compares this BFCM to last BFCM and reads a number that is no longer measured the same way.

What to report instead. These four are untouched by the change, so lead your BFCM readout with them:

• Orders

• Revenue

• Average order value

• Units sold

A conversion rate comparison needs both years measured the same way, which you no longer have inside Shopify.

Every BFCM readout we send this year leads with orders, revenue and AOV, and conversion rate comes with a footnote. If your conversion rate jumps 15% and orders do not move, that is not a win. Shopify changed the ruler. The worst outcome is a team getting credit or blame for a number that changed by itself.

Then mark the date. Shopify added annotations to analytics reports in this same release, so the note sits with the data. Use September 21, 2026, the date Shopify gives for the change.

Marketing reports in this release also separate traffic sources differently to cut misattribution, so channel-level session counts can move as well. If you report blended performance, meaning paid results measured against total site traffic rather than per channel, annotate those reports too.

Two of the three changes pull in opposite directions on accuracy. Consent-based counting drops real visitors who can still place orders. Bot filtering moves the number toward the truth, and the scale there is bigger than most stores expect. 

When we and Littledata measured bot traffic in the GA4 data of 787 online stores, 19% of the Littledata stores and 37% of ours recorded 5% or more of their sessions from bots, and every one of those sessions had already passed GA4's own bot filter.

 On the affected stores, real conversion rate ran 10% to 14% higher than GA4 reported. That study measured GA4 rather than Shopify analytics, so the figures are not your Shopify numbers, but the mechanism is the one Shopify has now changed.

Fix 2: check your published policies

Shopify can now generate return, contact, shipping and terms policies and keep them in sync. Turning that on replaces your existing text with Shopify's template, and Shopify's own documentation advises saving a copy of any custom wording and translations first.

So open all four policies and confirm they still say what you wrote, and leave the automation setting alone until after your freeze. A return policy that no longer matches what your support team is telling people turns into support tickets at your busiest point of the year.

The four standard policies on the Settings > Policies screen, each showing its published state.

That is the full pre-freeze list. Everything else in the release can wait, including the parts that look most appealing right now. The new abandoned-cart templates and the expanded campaign automation are both reasonable January projects.

Launching either during peak puts untested copy and timing in front of your highest-traffic week. Automated bidding needs a learning period, and BFCM is the most expensive week of the year to buy that learning. We would rather run what we know converts through Cyber Monday and give new automation a clean January baseline, when every wasted dollar costs less.

The 5 changes to handle after BFCM

Two of these have fixed dates, spread across five milestones:

Which of the two 2027 deadlines applies to public apps and to custom apps

Change 1: apps using script tags stop working on March 1, 2027

Shopify stops loading existing script tags on storefronts on that date. Any app injecting JavaScript this way loses that function unless its developer has moved to a theme app extension or a web pixel.

This one applies to any app using script tags, including a custom app that was built for your store.

Script tags are most common in apps that put something on the storefront: upsell and cross-sell widgets, review displays, tracking pixels, chat widgets and popups. If your stack has any of those, start with those vendors.

Change 2: apps on non-expiring tokens are delisted on January 1, 2027

Public apps still using non-expiring tokens are blocked from new installs on that date. Existing installs keep running until April 1, 2027. Custom apps and apps you built yourself are exempt from this one.

For public apps the developer does the migration, but you need to ask whether they have. Send this to every app vendor in your stack in December:

Two questions about [app name] ahead of the 2027 Shopify deadlines. Have you migrated to expiring offline access tokens before January 1, 2027? And does the app still use script tags, and if so what is the migration plan before March 1, 2027?

Chase anyone who has not replied by mid-January. If a vendor cannot answer either question in writing, start looking for a replacement.

Change 3: native features that replace paid apps

Three jobs merchants currently pay an app for are built in now:

• Scheduled stock counts

• Barcode scanning from any admin page

• URL-based digital products

If an app in your stack does one of those, January is when to check the bill. This is the most direct cost saving in the release.

Change 4: agent checkout and Shopify Catalog

Shop Pay now runs inside several AI assistants, and you can sell through Google Search AI Mode and Gemini, with orders arriving in your normal admin. Getting in depends on Shopify Catalog eligibility.

Your store needs to be:

• On the Starter plan or higher

• Out of password mode, with a verified email and two-step authentication on

• Current on Shopify invoices, with a low chargeback rate and no open policy issues

Each product needs:

• A title, at least one image and a price above zero

• Publication to the online store, Hydrogen or Headless

• To be visible, so nothing unlisted or hidden from search engines

Product field mapping for Shopify Catalog, showing how much of the catalog carries a title and a description, and the compare-at setting that exposes strikethrough pricing to AI agents.

Several categories are excluded outright, including alcohol, tobacco and vaping, CBD supplements, prescription and medical products, weapons, adult and sexual health products, event tickets, live animals, services and product add-ons such as warranties and subscriptions, and uncategorized products, meaning anything Shopify cannot classify from your product data.

One detail is easy to miss. Eligibility is reviewed over time rather than granted once, so an unpaid invoice or a rising chargeback rate can cost you the channel after you have it.

Shoppers already buy from AI referrals that land on your storefront, which is a different thing from agent checkout. Across 96 million sessions on 73 stores we manage, visits from AI assistants held flat year over year while orders from those visits rose to 2.74 times their level a year earlier. 

AI assistants are not sending more people. They are sending people who are closer to buying. That changes where to put effort in 2027. Less chasing AI visibility as a traffic number, more making sure the page someone lands on can be read and acted on.

Yarden Khoushy of Jurni worked through what that looks like in practice on Live with Shero, building a landing page around intent rather than volume.

Eligibility only gets your products considered. Our analysis of 1,000 Shopify stores put the average AI search readiness score at 42 out of 100:

• Product pages: 53

• Category pages: 35

51% of AI visits land on a product page, and those visits convert at 1.77%. Product pages are the stronger of the two, so category pages are where the work starts.

Change 5: rollouts and scheduled launches

You can now bundle theme, checkout, catalog and discount changes into one scheduled launch, run it across sales channels, and schedule promotional pricing that reverts on a date you set.

Set it up in January and next year's BFCM launch can be scheduled in advance.

One group to ignore

Several payment items carry Shopify's own coming soon or rolling out soon labels:

• Expanded fraud rules

• Chargeback protection beyond Shop Pay

• Custom checkout fields

• HSA and FSA payments

Revisit them when Shopify ships them.

Why the 2027 deadlines matter more than the new features

Everything else in this release is an improvement you can adopt at your own pace. The token and script tag dates are different, because they happen whether you plan for them or not.

There is a second reason to protect time in January. This release points toward shopping that happens inside AI assistants, and taking an agent order is only half of that problem. The other half is being picked in the first place, which depends on how well your product and category pages can be read. 

January is when most teams are recovering from peak, and that is exactly when two of these deadlines arrive. Protect two weeks of January for this, one for the app audit and one for AI readiness on your product and category pages. Neither will look urgent until it breaks.

If you want a score to start from, our Shopify AI readiness assessment covers schema, catalog data, content structure and channel eligibility.

F.A.Qs on Shopify Editions Fall 2026