Rollouts / New feature Admin

Rollouts
Schedule, stage, and A/B test themes and checkout

You can now schedule themes, checkout, and customer account settings to publish at a specific date and time, roll them out gradually to a subset of visitors, or A/B test two configurations as-is. The currently published theme/settings are automatically copied, so you can keep editing independently of the experiment.

On this page
  1. What can you actually do now (understand in 30 seconds)
  2. 5 ways to use it
  3. How it works: automatic creation of a published copy
  4. A visual of A/B testing and staged rollout
  5. Settings covered
  6. How to get started (3 steps)
  7. 5 key points for developers
  8. 3 use cases you can apply to your business
  9. A one-line summary for your pitch

1What can you actually do now

Shopify's Rollouts now supports, for theme, checkout, and customer account settings, "scheduled publishing," "temporary switching," "staged rollout," "A/B testing," and "market-specific localization testing".
It's a feature that lets you control changes—which used to go live in one shot—along a time axis and by traffic ratio.

Before: publishing was "all at once, right now"

Switching a theme or checkout setting took effect immediately for all visitors on the spot. You couldn't specify a date and time, publish to only a subset, or compare two options.

Rollouts: control by time and ratio

Specify "when," "to whom," and "at what percentage" to release. You can run two options at the same time and even measure which one wins. The currently published theme is protected by an automatic copy.

25 ways to use it

1. Scheduled publish

Switch everything at a specific date and time

Publish a new checkout or theme at a specified date and time. Example: on a certain date, switch from Dawn to Horizon.

2. Temporary switch

Swap in for a limited time → revert automatically

Temporarily swap to a different theme/checkout configuration. Example: enable a BFCM theme for just one week, then revert automatically.

3. Phased rollout

Release early to only a set percentage of visitors

Gradually roll out a new theme/checkout configuration to a % of visitors. Example: the new branded customer account experienceto a subset first.

4. A/B testing

Compare two configurations head-to-head

Run an A/B test between two entirely different theme/checkout settings to determine the winner. Example: a new upsell block at checkoutvalidation.

5. Market-specific localization testing

Schedule and test content localized per market

Schedule/test content localized per market. Example: different CTA text for each regionto serve variations and observe the impact on conversion.

3How it works: automatic generation of the published copy

Live theme Running live Production Auto-copy Duplicate of the live theme Continue editing here Experiment theme Deliver with Rollout Rollout Date & time / ratio / A·B control Assign to visitors
A copy of the live theme/configuration is generated automaticallyis the key point. While an experiment is running, you can keep editing the live theme and checkout independently of the experiment. Furthermore,mutually exclusive experimentscan be set up so that multiple live experiments run at the same time.

4A/B testing and phased rollout illustrated

A/B test: 50% / 50%

Visitors Variant A · 50% Variant B · 50%

Serve both variants simultaneously and determine the winner by conversion and other metrics.

Phased rollout: expand gradually

5% 25% 50% 100% Time →

If there are no issues, increase the ratio and eventually roll out to everyone.

5Settings covered

Theme
Switching between Dawn → Horizon, etc.
Checkout configuration
Upsell blocks, etc.
Customer account
Account page configuration
Detailed prerequisites such as supported plans, eligible countries, and usage limits are not stated in this release announcement. Before adopting it, check your store's eligibility in the Shopify admin under Markets > Rollouts and in the official Help Center.

6How to get started (3 steps)

1

Open Markets > Rollouts

From Markets in the Shopify admin, go to "Rollouts".

2

Create a Rollout

Create a new one with "Create a Rollout". Choose the target theme or checkout configuration.

3
A B

Set the method and conditions

Specify the scheduled date/time, rollout percentage, A/B test, markets, and more, then start.

The moment the experiment starts,a copy of the published theme/settings is created automatically,so you can keep editing the live version while running the experiment. For details, see the Rollouts documentation in the official Help Center.

75 key points for developers

1. "Publishing" now has a time dimension

Theme/checkout publishing, which used to be instant, now carries states like scheduled, phased, and time-limited. Rethink your operations calendar and deployment procedures on this premise.

2. Published copies prevent editing conflicts

When an experiment starts, a copy of the published theme/settings is generated automatically. Because live editing and the experiment branch are separated, you can keep applying live fixes even during the experiment.

A B

3. A/B testing compares "entirely separate configurations"

Rather than partial diffs, you can pit two completely different theme/checkout configurations against each other as-is. Well suited to comparing large redesign proposals.

4. Run concurrently with mutual exclusion

Designed as mutually exclusive experiments, you can run multiple live experiments at the same time. This enables multiple concurrent tests without overlapping test subjects.

Market

5. You can validate localization on a per-market basis

You can schedule and test content localized for each market (e.g., region-specific CTA text) and measure its impact on conversion. It can be built into the A/B design of multilingual, multi-currency stores.Handling via API/Webhook is not covered in this announcement,so if you rely on automation, separate validation is required.

83 use cases for your business

USE CASE 1

Make BFCM and major-sale theme switches unattended with "scheduling + automatic reversion"

Challenge
Manually publishing the theme late at night when the sale starts, then manually reverting it after it ends. This ties up a staff member and carries the risk of forgetting to revert.
Approach
Use a temporary swap to apply the BFCM theme for a specified period, and automatically revert to the original theme at the end time.
Result
Zero late-night work and zero forgotten reverts. You can run the launch and teardown of sale presentations on a calendar-driven basis.
Technical notes
Published copy is generated automatically, so you can keep editing the regular theme in parallel even during a sale.
A B
USE CASE 2

A/B test checkout upsell tactics with no implementation harness

Challenge
You want to add a new upsell block to checkout, but the impact is hard to predict, and building your own A/B measurement infrastructure is a heavy lift.
Approach
Run the current checkout and the "with upsell block" variant simultaneously via a Rollouts A/B test, and determine the winner.
Impact
You get decision-making data on whether to ship it—without building your own measurement infrastructure. It can also be repurposed as evidence in proposal materials.
Technical notes
If you set the experiments to be mutually exclusive, you can run them in parallel without overlapping subjects with other in-progress tests.
USE CASE 3

Test market-specific CTAs and localized copy by region for cross-border stores

Challenge
The messaging that resonates should differ by country, but you're shipping a single CTA and copy across all markets, so you don't know the optimal answer per region.
Approach
With per-market localization testing, serve different CTA text and content by region and compare the impact on conversion.
Impact
Optimize copy per market and explain the ROI of localization with numbers.
Technical notes
Combined with phased rollout, you can roll out a new brand experience to a subset of markets and visitors while keeping risk low.

9One-line summary you can use in a pitch

"Rollouts lets you take theme, checkout, and customer account settings and
schedule publishing / toggle time-limited changes / roll out in phases / run A/B tests / test localization by market."
Published copy is generated automatically so experiments don't clash with live edits, and mutually exclusive experiments can run several at once in parallel.
It lets you design 'when, to whom, and at what percentage to ship a big change' with no code."