Replatforming is the highest-stakes version of a website migration, because on an ecommerce site the organic traffic isn't an audience, it's revenue. A content site that loses a third of its traffic for two months has a bruised chart. A store that loses a third of its organic sessions loses a third of the sales those sessions were producing, every week, until it's fixed. I've written a general website migration checklist that covers the full sequence, and all of it applies here. This post covers the parts that are specific to moving a store, because ecommerce migrations fail in their own particular ways.
Why store migrations are harder than they look
Three things make replatforming riskier than an ordinary rebuild.
The first is scale. A store has a URL for every product, every variant the platform exposes, every category, every filtered view of every category, and every page of pagination behind them. A ten-page brochure site can have its redirect map written by hand in an afternoon. A catalogue of 4,000 products cannot, and the temptation to cut corners on mapping grows with every thousand URLs.
The second is that the new platform usually forces URL changes whether you want them or not. This isn't a design decision you can push back on; it's baked into the software. Moving to Shopify means your category URLs now live under /collections/ and products under /products/, with no way to opt out. That single constraint means a migration to Shopify changes every commercial URL on the site, which makes the redirect map the entire job. I've covered the platform's quirks more fully in the Shopify SEO guide.
The third is that your rankings are concentrated in pages that are easy to break. On most stores the category pages carry the commercial rankings, and category pages are exactly what gets restructured, merged, or renamed during a replatform, usually by someone thinking about navigation rather than search equity.
What each platform forces on your URLs
Worth knowing before you choose a destination, because the amount of redirect work varies significantly by platform.
| Platform | URL structure | What that means for the migration |
|---|---|---|
| Shopify | Fixed: /collections/, /products/, /pages/ | Every commercial URL changes. Full redirect map, no exceptions |
| WooCommerce | Fully customisable | You can often replicate old paths exactly and skip most redirects |
| BigCommerce | Customisable | Same, with some platform-specific defaults to override |
| Magento / Adobe Commerce | Fully customisable | Replication possible, but complexity means it's often skipped anyway |
If you're moving between two flexible platforms, replicating the existing URL structure is nearly always the right call. Every URL you keep identical is a redirect you don't need, a hop that never happens, and a page Google doesn't have to re-evaluate. The clean-slate instinct ("while we're at it, let's tidy up the URLs") converts a low-risk migration into a high-risk one for cosmetic benefit.
The redirect map is where the revenue lives
For a store, I build the redirect map from revenue downwards rather than from the sitemap downwards. Export every URL from analytics with its organic revenue attached, every URL with external links, and every URL from a full crawl. The pages that make money and the pages that hold authority get mapped first and checked by a human. The long tail can be mapped by rule (pattern matching old paths to new), but the top hundred URLs deserve individual attention, because a mapping error on a category doing £5,000 a month of organic revenue costs more than the entire migration budget saved by rushing.
Map one hop, to the closest equivalent page, never in bulk to the homepage. Discontinued products go to their nearest category, not to the root. And carry forward the redirect rules from any previous migration, because a store that's been through two replatforms has redirect debt from both, and dropping the legacy rules quietly severs every old link pointing at the site. Platform limits are rarely the problem in practice; Shopify supports 100,000 native redirects on standard plans and far more on Plus, which is enough for almost any catalogue if the long tail is handled by pattern rather than line-by-line.
Change one variable at a time
The most expensive mistake I see is bundling the replatform with a category restructure, a rebrand, and a redesign in one launch. Each of those is a variable that can move rankings on its own. Launch all four at once and, when traffic drops, nobody can say which one caused it, so nobody can fix it.
If the category structure genuinely needs rework, do it as a separate phase, either on the old platform before the move or on the new platform a couple of months after, once rankings have settled. The same goes for faceted navigation, which the new platform will handle differently from the old one by default. Decide deliberately which filtered views should be crawlable and which shouldn't, because the default settings on the new platform were not chosen with your index in mind.
After launch, watch the money pages
The general checklist covers verification and monitoring, so here I'll just add the ecommerce-specific habit. Benchmark rankings and organic revenue for your top categories and products before launch, then track those specific pages daily for the first few weeks rather than watching the site-wide traffic line. A site-wide chart can look flat while one high-revenue category has dropped and a cluster of blog posts has picked up the slack, and you want to catch that in week one, not at the quarterly review. Some movement is normal while Google re-evaluates changed URLs; what isn't normal is a money page still down after four to six weeks, which nearly always traces back to a redirect gap, a template problem, or content that got thinner in the move.
This is the kind of work I do as part of ecommerce SEO engagements, and the honest advice is that the cost of doing a migration properly is a fraction of the revenue at stake in getting it wrong.
Ecommerce migration FAQs
Will we lose traffic even if we do everything right?
Expect some turbulence. Google has to recrawl and re-evaluate every changed URL, and on a large catalogue that takes weeks. A well-run migration typically sees a modest dip that recovers within one to two months. What separates a dip from a disaster is whether the redirect map was complete and the templates preserved the content that was ranking.
Should we redesign at the same time as replatforming?
It's usually unavoidable, since moving platform means new templates. What you can avoid is adding a category restructure or a rebrand on top. Keep the information architecture and the content as stable as you can through the platform move, then iterate on the new platform once rankings have settled.
Do discontinued products need redirects?
If the URL has external links, ranks for anything, or still gets traffic, yes, redirect it to the closest category or replacement product. Pages with none of those can be left to return 404, which is a perfectly clean signal. What you shouldn't do is redirect thousands of dead products to the homepage, which Google treats as a soft 404 and ignores.
How long should we keep the old redirects?
Indefinitely. Redirect rules cost nothing to keep, and any old URL with backlinks keeps passing equity only while the redirect exists. The common failure isn't keeping redirects too long, it's losing them in the next replatform because nobody carried the rules across.
Planning a replatform?
Tell me what you're working on and what you're trying to achieve, and I'll give you an honest view of whether I can help and what it would take.
Get in touch