What comes across and what gets left behind, why URLs change shape and how to redirect them, what to test before you cut over, and the thirty days after.
by Nadia Maar ·
Migrations don't fail on the catalogue: that exports and imports. They fail on URLs. A store with a few years behind it has hundreds of indexed pages, and on the new platform those pages will have different addresses: if nobody connects them, organic traffic slides over the following weeks, and you notice once it has already happened.
This is the map: what comes across, how to redirect, what to test before cutting over, and what to watch in the thirty days after.
| Comes across | Stays behind, or gets rebuilt |
|---|---|
| Products, variants, prices, images | Customer passwords: accounts have to be reactivated, and the email asking for it needs writing well |
| Customer records and addresses | Theme customisations and plugins: rebuilt with the new platform's tools |
| Order history, usually as an archive | Reviews, if they live in a plugin: you need an export, and one doesn't always exist |
| Blog posts and pages | Forms and automations wired to the old site |
The password row generates the most support after launch: customers don't understand why their login stopped working. A clear email, sent on switch day and not three days later, solves it.
The two platforms organise addresses differently: WooCommerce usually uses a form with "product" and "product-category", Shopify uses "products" and "collections". It looks like a detail, and it's the difference between a successful migration and a lost quarter.
Redirect from,Redirect to
/product/leather-bag,/products/leather-bag
/product-category/bags,/collections/bags
/shop,/collections/all
/cart,/cartShopify imports redirects from a file with these two columns, and accepts wildcard rules for paths that follow a pattern. General rules cover the bulk; the pages that bring traffic get checked one by one.
A correctly imported product can still lose rankings if everything around it is lost:
A dip in the first two weeks is normal: engines have to re-read the site and follow the redirects. What isn't normal is failing to notice where you're losing. Four things to watch, in this order:
If after a month organic traffic hasn't come back close to where it was, the problem isn't time: it's the URL map. You re-read it, complete it, and recover.
They're always the same ones, in every migration that arrives needing repair. None of them is hard to avoid: they're hard to remember while you're thinking about everything else.
And a sixth, which isn't a technical mistake: promising the sales team that "nothing changes". Internal URLs change, field names change, the way you look up an order changes. Half an hour of explanation before the switch is worth more than two days of phone calls after it.
The work that protects traffic — the URL inventory, the redirect map, the testing — happens in the preceding weeks. Afterwards you can only chase.
If you're weighing the move, the free audit starts from the inventory: what brings visits and orders today, and what it takes not to lose it.
Only if the URLs aren't mapped. The platforms use different shapes — "product" and "product-category" versus "products" and "collections" — and every old URL has to redirect to its equivalent, not to the homepage. A two-week dip is normal; a dip that stays means an incomplete map.
Customer passwords, which have to be reactivated with a well-written email; theme customisations and plugins, rebuilt with the new platform's tools; reviews, if they live in a plugin with no export.
A real order on the new store, actually paid and shipped; redirects verified against the list of pages that bring traffic; DNS time-to-live lowered a few days earlier; the old store still reachable.
Crawl errors and not-found pages, the position of the twenty pages that brought the most traffic, checkout conversion, and support requests. In that order.