How to change ecommerce platforms without losing SEO traffic

Replatforming is one of the few projects where a store can do everything right commercially and still go backwards. The new site is faster, the admin is better, the theme finally looks like the brand, and then organic traffic falls by a third and does not come back. The cause is almost never the new platform. It is that the store treated the migration as a rebuild of the front end, when search engines experience it as the sudden disappearance of several hundred addresses they had spent years learning to trust.

The useful mental model is this: your rankings are not attached to your store. They are attached to specific URLs. Every link another site has given you, every signal accumulated from people clicking your result, all of it is filed against an address. Change the address without leaving a forwarding note and the file is orphaned. Leave a correct, permanent forwarding note and the vast majority of that value transfers. Almost everything else in a migration plan is detail hanging off that single principle.

What usually transfers	What usually breaks
Ranking signals, when URLs are correctly redirected	URLs, because platforms use different structures by default
Product and category copy, if exported in full	Truncated descriptions and lost meta titles or descriptions
Brand and direct traffic	Blog and resource pages, often forgotten in the export
Reviews, if migrated with an app or export	Structured data, review markup and image alt text
Domain authority, provided the domain itself is unchanged	Analytics continuity and conversion tracking

Why does organic traffic drop after a replatform?

Four failure modes account for most of the damage, and they compound. The first is unmapped URLs. Platforms impose their own patterns, so a product that lived at one path now lives at another, and every old address quietly returns a 404. The second is lazy redirects, usually sending every old page to the homepage. That technically avoids the 404, but it tells a search engine the specific page no longer has an equivalent, and the ranking signals dissolve rather than transfer.

The third is content loss. Exports frequently truncate long product descriptions, drop custom meta titles, and lose image alt text. A page that ranked on 600 words of genuinely useful detail will not hold its position on the 90-word summary that survived the export. The fourth is launching a site that blocks crawlers. Development environments are usually set to discourage indexing, and if that setting travels to production, the new store is invisible until someone notices. Each is preventable, and none is easy to reverse afterwards.

What should you do before you touch the new platform?

The work that protects your traffic happens before anything is built. Start by taking a full inventory of what you actually have. Export every URL that currently exists on the site, including product, collection, blog, policy and any legacy pages that still attract visits. Pull the same list from your analytics and from Google Search Console, because each source knows about pages the others have forgotten, and the pages you forget are exactly the ones nobody thinks to redirect.

Then record a baseline. Note current organic sessions, revenue from organic, the pages that generate the most traffic, and the queries you rank for. Without a baseline you cannot tell a normal post-launch dip from a real problem. Prioritise ruthlessly: in most catalogues a small share of pages drives the majority of organic revenue. Those pages get individual attention. The long tail gets rule-based handling.

This is also the moment to be honest about whether you are solving a real constraint or a frustration. If you are still weighing the options, the trade-offs are worth working through first in Shopify vs WooCommerce for small business, because the cheapest migration is the one you did not need.

How do you map old URLs to new ones?

The URL map is the core deliverable of the whole project. It is a simple two-column list: every old URL, and the single best matching new URL. Build it in a spreadsheet and have a second person review the top pages. Where a like-for-like match exists, use it. Where it does not, redirect to the closest genuine equivalent, usually the parent category. Only pages with no reasonable equivalent should be allowed to return a 404, and that should be a deliberate decision rather than an oversight.

Old page type	Redirect target	Common mistake
Product, still sold	The same product on the new platform	Redirecting to the category instead
Product, discontinued	Nearest replacement or its category	Sending it to the homepage
Category or collection	Equivalent new collection	Losing filtered or sorted variants
Blog or guide	Same article at its new path	Excluding the blog from the migration entirely
Policy and shipping pages	Equivalent new page	Assuming nobody links to them

Implement the map as 301 redirects, which signal a permanent move and pass ranking signals to the destination. Avoid 302s, which signal a temporary move and hold the signals in place. Check that each redirect resolves in one hop: a chain of three or four redirects wastes crawl budget and slows the transfer. Keep the redirects in place indefinitely. They are cheap to maintain and expensive to remove.

What has to move with the page, not just the URL?

A redirect points at a page. What matters next is whether the page it points at deserves the ranking. Content parity means the new page carries the same substance as the old one: the full product description rather than an abbreviated version, the same title and meta description unless you have a reason to improve them, the same headings, and image alt text that survived the transfer. If a page ranked on depth, publishing a thinner version and hoping the redirect carries it is wishful thinking.

Alongside content, check the technical furniture. Canonical tags should point to the live new URL, not to a staging domain, which is one of the more common and more damaging launch bugs. Structured data for products, reviews and breadcrumbs should be regenerated on the new platform rather than assumed. Internal links inside body copy should be updated to the new paths so they do not rely on redirects forever. And because the new theme is also a new shopfront, it is worth confirming the pages still answer buyer questions properly, which is the subject of improving product page conversion rate. Traffic you keep but no longer convert is only half a win.

How should you sequence launch?

Migrations fail on timing as often as on technique. A defensible sequence looks like this:

• Build and stage privately. Keep the staging site blocked from indexing so it never competes with the live store.
• Load the redirect map before you switch. The redirects should be live the moment the new site is, not added the following week.
• Unblock crawling at launch. Explicitly confirm the production site allows indexing. This single check prevents the worst outcome.
• Launch in a quiet period. Choose a low-season window, not the run-up to a peak trading period, so an unexpected problem costs the least.
• Submit the new sitemap and monitor crawl errors daily for the first fortnight.
• Verify tracking for analytics, conversions and ad platforms before you rely on any number the new site reports.

Resist the temptation to replatform, redesign and rewrite at once. It is understandable, but if traffic and conversion both shift and four variables changed together, you cannot tell which one to fix. Migrate with content parity, stabilise, then improve deliberately. If local visibility matters, re-check the fundamentals in getting your local business found on Google once the dust has settled.

How long does recovery take, and what should you watch?

Some fluctuation is normal: search engines have to recrawl the site, follow the redirects and reassign signals, and none of that is instant. For a clean migration on a small to mid-sized catalogue, most of the recovery commonly lands within four to eight weeks. Larger catalogues take longer simply because they are crawled more slowly. A modest dip in the first fortnight is expected; a steep fall that keeps falling is a fault, not a phase.

Watch four things weekly against your baseline: organic sessions, indexed page count, crawl errors, and organic revenue. If pages are dropping out of the index, look for blocked crawling or canonical tags pointing somewhere unintended. If a specific high-value page collapses while others hold, check its individual redirect and whether its content came across in full. If traffic remains materially below baseline after about three months, stop waiting and audit the redirect map properly. Diagnosing this pattern is part of what our e-commerce and digital transformation work covers.

The decision behind the decision

Changing platforms is not really a technical decision, it is a bet that the new platform removes a constraint worth the disruption. When that constraint is genuine, the migration is a manageable, temporary cost with a clear method for containing it. When the reason is mostly aesthetic, you are spending real risk to buy a nicer admin panel. Know which one you are doing, then execute with the discipline the first case deserves. The stores that come through a replatform stronger are not the ones with the best new theme. They are the ones that treated their URLs as an asset worth protecting, mapped them carefully, and checked their work. If you are planning a move and want the mapping and launch sequence reviewed before you commit, book a consultation and we will work through it with you.