Shopify Rollouts can schedule, stage, or experiment with storefront changes; use them with a defined customer path, metric, guardrail, and rollback owner.[1][5]
A theme change is not only a design decision. A new product-page layout can alter how shoppers find size information, a beauty routine, delivery details, reviews, or the add-to-cart control. A seasonal storefront can also change what paid visitors, product feeds, and returning customers see. Rollouts gives a Shopify team a controlled place to manage that change, but it does not decide whether the change is commercially safe.
Shopify provides three rollout types. The type determines whether your change is permanent, temporary, or compared with a control, as well as which analytics and audience settings apply.[5]
| Rollout type | Use it for | Do not use it as |
|---|---|---|
| Launch | A change you have already reviewed and want to apply permanently, either immediately or with a planned reach. | A substitute for evidence when the team is still deciding whether the change works. |
| Event | A time-bounded storefront, checkout, catalog, or discount change that should roll back when the event ends. | A permanent redesign with no owner for the final state. |
| Experiment | A treatment-versus-control comparison when you have a clear hypothesis and a way to interpret the result. | A redesign that changes so many customer jobs that no result can explain what happened. |
For example, a fashion brand might use an Event for a seasonal campaign theme, an Experiment for a revised size-and-fit section, and a Launch for a product-page structure that has already passed review. A beauty brand could test a routine-led product-page layout against its current ingredient-led layout, while a broader Shopify merchant could compare two ways of presenting shipping and returns information.
Write the job in one sentence before you open the admin: “We are using a ___ to ___ for ___.” If the blank is “find out whether,” you are describing an Experiment. If the blank is “show this during a defined promotion,” you are describing an Event.
Rollouts are available on Basic and higher plans, but Shopify says Experiments require Grow or higher. Staff also need the Rollouts permission and additional permissions for the resource being changed, such as Online store permissions for a theme replacement or Edit code permission to save theme-editor changes to a rollout.[2]
Check the boundaries before writing a test brief:
These limits are not reasons to abandon a storefront improvement. They tell you whether the change belongs in Rollouts, needs a separate implementation path, or should be released without pretending it is an experiment.
A useful hypothesis names the customer problem, the treatment, and the expected decision. “The new theme looks cleaner” is a review comment. “Mobile shoppers cannot find delivery timing before adding a product, so moving a verified delivery message beside the purchase control may improve checkout reach without increasing support questions” is a test brief.
Use evidence your team already owns to choose the problem:
Keep the treatment narrow enough that the result teaches you something. Shopify describes A/B testing as a controlled comparison and recommends beginning with one change at a time rather than changing several independent elements together.[4] A full theme replacement can still be a valid Experiment, but state what the replacement is intended to improve and what you will inspect if the main metric moves.
Complete these fields before you create the rollout:
| Decision field | What to record | Example |
|---|---|---|
| Customer problem | The observed question or friction | Shoppers cannot find fit guidance before choosing a size. |
| Treatment | The exact theme or configuration change | Put the verified fit guide beside the size selector and keep the existing product details below it. |
| Audience and market | Eligible market, storefront, and reach | US online store visitors on the fashion market. |
| Primary metric | One outcome that matches the customer job | Add-to-cart rate or reached-checkout rate. |
| Guardrails | What must not degrade | Product price, inventory, checkout errors, page speed, returns context, and support load. |
| Decision owner | Who reviews the result and chooses the next action | Ecommerce lead with merchandising and analytics partners. |
| End action | What happens after the measurement period | Apply, keep control, narrow the treatment, or stop. |
| Recovery path | What can be paused, reverted, or re-applied | Named owner, saved control, and documented customer-path checks. |
The main measure is a decision aid, not a promise. If the change is intended to help product discovery, an add-to-cart or product-engagement measure may be closer to the question than completed orders. If the change affects checkout, use a checkout measure. Pair the decision measure with risks to monitor so a result that looks positive in one report does not hide slower pages, more returns, broken discounts, or additional support work.
Shopify’s current creation flow starts at Markets → Rollouts → Create rollout. You name the rollout, select Launch, Event, or Experiment, add the change, and configure the audience and traffic allocation.[6]
For an Experiment, Shopify provides a control-and-treatment traffic split; the creation guidance describes a new Experiment as using a 50% control and 50% treatment default, which you can adjust for your store’s risk and evidence plan.[6] Do not confuse the configured allocation with the number of eligible visitors who actually saw each experience when other markets or rollouts affect reach.[2][3]
Before sending visitors to the treatment, review the complete customer path:
A visual preview is not enough. A fashion product can look correct while its size selector fails. A beauty product can render correctly while an ingredient or usage section disappears on mobile. A seasonal banner can look persuasive while pointing paid traffic to an expired offer. Test the actions the merchant needs the storefront to support.
Rollout analytics depend on the type and number of changes, and Shopify says the available metrics cannot be customized. Theme experiments can expose measures such as conversion rate over time, bounce rate, reached checkout rate, and added-to-cart rate; the exact metric set depends on the rollout configuration.[3]
Use the result in layers:
Do not report a theme preference as a conversion win. Do not treat a change observed during a promotion as a clean test of the theme alone. Shopify also notes that the same change can produce different results in a Launch and an Experiment because the rollout types use different analytics methods.[3]
Keep the treatment and control stable while an Experiment is active. Shopify warns that editing either side can affect the results.[7] If a defect requires a change, record what changed, when it changed, which visitors were exposed, and whether the original decision is still interpretable.
Use a bounded decision rather than a forced winner:
A Launch merges its changes permanently, while an Event is designed for a defined period and rollback. An Experiment exists to compare a treatment with a control before a broader application.[5] Name the end action before launch so the team is not making a permanent decision because a dashboard is convenient.
Before you publish, confirm:
Rollouts is most useful when it makes a merchant decision more explicit: what is changing, for whom, why it should help, how the team will know, what must not get worse, and who acts next. If those answers are missing, improve the brief before changing the storefront.
[1] https://help.shopify.com/en/manual/markets/rollouts — Shopify Help: Rollouts [2] https://help.shopify.com/en/manual/markets/rollouts/requirements-and-considerations — Shopify Help: Rollouts requirements and considerations [3] https://help.shopify.com/en/manual/markets/rollouts/analytics — Shopify Help: Rollout analytics [4] https://www.shopify.com/blog/ab-testing — Shopify: A/B Testing [5] https://help.shopify.com/en/manual/markets/rollouts/rollout-types — Shopify Help: Types of rollouts and changes [6] https://help.shopify.com/en/manual/markets/rollouts/create-a-rollout — Shopify Help: Creating a rollout [7] https://help.shopify.com/en/manual/markets/rollouts/managing-rollouts — Shopify Help: Managing rollouts