Moving to Shopify Without Losing Orders: A Go-Live Runbook

Direct answer: moving to Shopify without interrupting sales is an operational objective, not a guarantee. The safest launch keeps the source store authoritative until a defined cutover, reconciles changes made since the first migration, tests the entire purchase journey and assigns clear stop or rollback criteria. A successful deployment alone does not prove that new orders can be processed.

The plan below is for a typical ecommerce migration. Actual timing, supported imports and handover responsibilities depend on the current platform and the systems connected to it.

Keep the existing shop selling while you build

Prepare Shopify separately from the customer-facing store. Configure the product catalogue, variants, collections, content, theme, payment options, shipping rules and integrations before switching the main domain. The old store can continue to accept orders during preparation, provided the team has decided which system is responsible for inventory and fulfilment at each stage.

Two independently editable catalogues create a risk: a product sold on the legacy store might still appear available in Shopify. Avoid treating a one-time CSV import as a live synchronisation. Document whether stock, prices and customer changes will be captured again before launch, and identify the system of record for each field.

Shopify's official migration guide provides the setup sequence, but your operational runbook must also describe the particular source platform and connected services.

Define the cutover window before touching DNS

A cutover is a sequence of decisions, not the moment a theme is published. Agree on a launch owner, a communications channel, a change freeze, a checklist and the evidence that allows each step to proceed.

Distinguish changes that are easy to reverse, such as a content edit, from changes that affect active orders, domain routing or external integrations. Record who can restore each component if an acceptance test fails. Schedule the handover when the relevant merchant, fulfilment and technical contacts can respond, rather than assuming that a quiet traffic period eliminates risk.

For each checkpoint, state an explicit go/no-go rule. If payment processing cannot be verified, or a required fulfilment system cannot receive orders, don't proceed because the homepage looks correct.

Reconcile the orders and stock created since the first import

The first export usually becomes stale. Between rehearsal and launch, the old store may receive new purchases, cancellations, refunds, customer registrations and inventory adjustments. Define exactly how that change set will be recovered.

Depending on the source platform, use a supported incremental export, a carefully scoped integration or a final reconciliation window. Preserve original identifiers and distinguish a previously imported record from a new transaction. Re-importing every order without deduplication can make financial and operational reports unreliable.

Check inventory at the variant level, especially if multiple locations or fulfilment services exist. A recent stock count must not be overwritten with an older export just because the CSV was prepared earlier.

For customers and historical purchases, see the customer and order history migration guide. Historical data and orders placed during the cutover require different acceptance tests.

Test checkout and what happens after payment

A store is not ready merely because products and prices render. Test a complete purchase flow using realistic combinations of product variants, discount codes, shipping addresses, taxes and payment outcomes. Verify that the resulting order appears with the expected status and that notifications, fulfilment and connected systems receive it correctly.

Shopify's test-order documentation explains supported test methods. An important limitation: putting a live payment provider in test mode prevents customers from placing genuine card orders through it. Keep test-mode use confined to the appropriate setup window, and turn it off before accepting live orders. A real payment followed by a refund can incur fees, so it should not be treated as a free test.

Where relevant, verify ERP integration, shipping labels, invoice generation, email templates and refund workflows. If a test checkout succeeds but an essential warehouse system never receives the order, the launch criterion has failed.

Switch the domain and validate old URLs

Before the DNS change, confirm the domain configuration, HTTPS certificate, www versus non-www behaviour and the intended canonical host. Prepare a mapping from important legacy URLs to their relevant new destinations, including product pages, collections and editorial content.

Google's site move guidance recommends matching old pages with relevant new pages and using suitable permanent redirects. Sending all previous URLs to the new homepage is not a substitute for a migration map.

Once the cutover starts, test real routes on the public hostname. The homepage can work while a product with incoming links returns a 404, the cart fails or a redirect points to the wrong language. Domain propagation alone does not demonstrate complete availability.

Make rollback operational, not theoretical

Write down the conditions that trigger a stop and the actions each responsible person can take. A minor spacing problem should not be treated in the same way as a broken checkout, incorrect taxes, overselling or missing order transmission.

A rollback may redirect customers to the old storefront or restore a previous integration configuration. It cannot erase Shopify orders that have already been accepted. Decide in advance how to identify and reconcile any orders created during a partial launch to avoid double fulfilment, duplicated payments or missing customer communications.

Keep access to the legacy data, necessary backups and the original URL map. Do not improvise a full restoration after you discover that the old store's inventory has already changed.

A practical launch-day acceptance matrix

Use a short, traceable runbook. These checkpoints describe evidence to collect, not a guaranteed timetable.

Checkpoint Evidence required before proceeding
Before cutover Catalogue, prices, shipping and payment settings signed off
Final data delta New orders and variant-level stock reconciled without duplicates
Public domain HTTPS, canonical host and priority redirects working
First real order Checkout succeeded and downstream systems received the order
Post-launch review Orders, failed payments, stock, email and important 404s monitored

The team should also verify a failed-payment path, a refund scenario and a returning-customer journey where these are material to the business. Keep a simple incident log recording symptoms, owner, action and verification outcome.

When to stop instead of pushing through

Stop if prices or stock no longer reconcile, if an essential payment method fails, or if fulfilment has no trustworthy record of new orders. The correct response depends on the issue: some failures can be fixed without reversing the whole launch; others require returning traffic to the previous checkout while reconciling transactions.

A green build, passing preview tests and an accessible homepage each provide useful evidence, but none replaces checking a genuine public order path and the associated business processes.

Continuity comes from preparation and verification

The objective is to reduce and contain disruption, not to promise zero risk. A documented cutover, controlled final data transfer, working redirects and clear recovery responsibilities make the transition testable.

For a project that covers the destination storefront and its operational dependencies, the Shopify migration service is handled directly by Francesco Guiducci, from migration scope to launch verification.

Sources cited in this article

Official documentation and external references linked in the text. Shopify features change: always check against the current version.

With your consent, Google Analytics 4 measures visits and interactions. With both categories enabled, it also uses Google click identifiers to attribute requests to Google Ads. Marketing also enables Meta Pixel for page views and WhatsApp clicks. Google and Meta do not receive form entries, messages or contact details. Google remarketing and ad personalisation stay off. You can refuse or withdraw from the footer. With Marketing enabled, the OpenAI pixel measures visits and may detect contact details from forms and send them as hashes with advertising identifiers.