Shopify Markets vs multiple stores: how to choose
Decide whether to localise one store with Shopify Markets or run independent stores for different operations.
Read articleShopify Rollouts: schedule changes and test them thoughtfully
Schedule campaigns or compare changes without confusing a real experiment with a simulation or a backup.
Shopify Rollouts lets you prepare and schedule sets of store changes as a permanent launch, temporary event or experiment. Rollouts are available from Basic; Experiments, comparing variants with visitors, require Grow or higher. Each resource and customisation retains its own requirements.
The tool organises change but does not remove risk. Decide what to modify, for whom, how to verify results and what to do if something fails. Scheduling does not replace those decisions.
Launch, Event and Experiment do different jobs
| Type | When it makes sense | What not to infer |
|---|---|---|
| Launch | Permanently publish a set of changes. | A rollout is not automatically an A/B test. |
| Event | Apply changes for a period and end their effect afterwards. | It does not restore all business data to a past date. |
| Experiment | Compare treatment and control with eligible traffic. | It guarantees neither improvement nor a conclusion with few visits. |
The rollout types guide describes supported resources, including themes, checkout and accounts, catalogues and discounts. Not every code or data change can be included identically. Shopify states, for example, that Liquid templates are not modified as part of a rollout.
Prepare a campaign event
- Define what must activate: presentation, commercial conditions and available products.
- Check the scheduled time and time zone.
- Review dependencies with campaigns, emails, feeds and support.
- Test the full journey under the offer's conditions.
- Verify what ends when the Event closes and which data remains.
- Assign someone to supervise the start and finish and retain a result reference.
A catalogue can affect product availability and pricing across channels depending on configuration. Ending an event does not reconstruct a historical catalogue backup: changes made to that resource during the period may remain. Also review feed and POS synchronisation times.
Design an experiment that produces learning
Form a specific hypothesis: observed friction, the change expected to resolve it and the metric testing its effect. A button label, information hierarchy and promotion are different tests; combining them prevents identifying which explains results.
Define the primary metric, harm signals and planned duration before starting. Traffic and the effect size you need to detect determine whether a useful conclusion is possible. For low-volume stores, investigating journeys and testing qualitatively may be more reasonable than declaring a winner from a few orders.
- Do not change a variant while gathering results except to fix an incident, and document it if you do.
- Avoid interpreting one day's spike as sustained improvement.
- Check whether promotions or channels explain the difference.
- Separate real visitor results from AI shopper simulations.
Channel, permission and customisation limits
Experiments and rollouts shown to less than one hundred percent of visitors apply to the online store. An offer may differ from a feed or another channel, so review promotional consistency. Market customisations have additional plan conditions.
Assign permissions by change type. A content editor does not necessarily need full code access or the ability to alter commercial conditions. Before scheduling, identify the resource and its owner.
A rollback plan that accurately describes its scope
Document the starting configuration and how to stop exposure to a change. Retain shared data and resources that may change outside the rollout. Ending an event does not undo orders, sent emails or external integration operations.
After launch, verify effective configuration matches expectations. Scheduling reduces manual activation work; quality depends on preparing a correct offer, measuring properly and reviewing results.
Continue working on your store
Prepare a high-demand campaign · Review performance before publishing · Automate operational tasks with control.
If you need to turn these options into a work plan, our Shopify consulting reviews your operations and dependencies before recommending an implementation. Tell us what your store needs.
Sources and review
Editorial review: the Migraciones.io team · 30 October 2026. Features, plans and availability are checked against the linked sources.