A store migration is a content migration with money on the line. The general rule still governs — nothing an organization owns gets lost — but commerce adds its own manifest, plus one item that genuinely cannot make the trip and needs to be planned around rather than promised.
What must survive
Order history. Years of records that your accounting, your customer service, and sometimes your legal obligations depend on. “The old orders will stay on the old system” is a red flag — the old system is going away; that’s the whole point.
Customer accounts. Emails, addresses, and history. Passwords often can’t transfer between platforms (different systems scramble them differently — a good thing), so the honest pattern is a planned password-reset email, announced, not discovered.
Products with their real complexity. SKUs, variants, categories, images, and the descriptions someone spent years writing.
URLs. Product pages accumulate search ranking and inbound links like any content — every old URL redirects to its new home, permanently.
What can’t come along: saved cards
Stored payment methods live with your payment processor inside PCI scope, and they don’t simply export. Whether they can carry over at all depends on the processor pair — same processor on both ends is often migratable; switching processors usually means customers re-enter cards on their next order. This is a planning item, not a failure: stores that know it going in communicate it and barely feel it. Stores that discover it at launch feel it plenty.
The cutover
Revenue hates downtime, so the pattern is: build and test the new store in parallel, place real test transactions on it, then cut over in a quiet window with a short, announced ordering freeze — hours, not days. Orders placed during the freeze window are the easiest thing to lose track of, so the freeze is honest and brief. Like a portage between lakes: the carrying is fine — it’s the put-in and take-out where things get dropped, so that’s where the care goes.
If the store that needs migrating is also the store nobody maintains anymore, stabilize checkout first, then migrate. Order matters.