A Shopify migration that handles SEO before the switch

A redesign often changes several layers at once: URL structure, navigation, content, internal linking, structured data, load times and how templates render. If these changes are only looked at near the end, the SEO team inherits an already-frozen site and an incomplete redirect list.

We bring SEO owners into the decisions that affect crawling and indexing. A deleted collection needs a destination or a documented reason. A new taxonomy needs to connect to the real catalog. A product template needs to expose the expected information and tags. Consolidated content needs to preserve useful links.

The SEO plan becomes part of the migration blueprint this way. It isn't a patch bolted on the night before launch.

The source crawl gives a first view, but it isn't enough on its own. It needs to be cross-checked against sitemaps, analytics data, Google Search Console, available backlinks, the catalog, and pages the teams already know about.

The goal is to build a URL inventory with status, type, traffic where that data is available, indexability, a likely destination, and an owner for each one. Orphan pages, parameters, old campaigns, and files all need to be identified. Variants in protocol, subdomain, or trailing slash can reveal duplication that already exists before the migration even starts.

A complete inventory doesn't mean every URL gets kept. It lets you decide without confusing a deliberate removal with something simply forgotten.

Not all URLs play the same role. An active product page, an editorial collection, a buying guide, an old campaign, and an account page don't call for the same treatment.

We classify URLs by type and by decision: keep the address, redirect to an equivalent, consolidate into a more useful page, delete with the right status, or exclude because it isn't indexable anyway. Priority follows the available signals, not a one-size-fits-all threshold.

Classification makes review easier. Merchandising can validate the collections. Content can confirm the consolidations. SEO can analyze traffic, queries, and links. Engineering can check that the destination actually exists in the target model.

A useful redirect connects a source intent to the best target destination. Redirecting every old page to the homepage hides errors and degrades the experience. Creating redirect chains adds unnecessary detours for both bots and users.

The matrix needs to contain, at minimum, the source URL, its current status, the destination, the rule type, the justification, and the planned check. It flags conflicts, loops, chains, and missing destinations. Generic rules can handle URL families when their behavior is genuinely uniform. Important pages deserve individual review.

Implementation depends on Shopify's capabilities and the architecture chosen. Limitations need to be tested before launch, not discovered after the import.

A redirect doesn't replace a clean internal link. Once destinations are defined, we update the links in navigation, collections, articles, editorial pages, and reusable components. This avoids leaving links that permanently pass through a redirect.

The new architecture also needs to stay understandable. Strategic pages need coherent navigation paths, links from relevant content, and a clear place in the hierarchy. A migration can be a good moment to fix weak internal linking, but changes need to be documented to tell a deliberate improvement apart from an accidental disappearance.

Technical SEO depends on what the templates actually output. We check titles, descriptions, headings, canonicals, robots directives, structured data, pagination, hreflang, HTTP statuses, links, and sitemaps, depending on scope.

Tests need to cover several instances of each type: an available product, an unavailable product, a product with variants, an empty or paginated collection, an article, a market page, and search results where relevant. A single product page won't reveal issues tied to specific data or states.

Staging environments need particular attention. Their indexing block must never get copied by mistake onto the live site. Conversely, a test environment shouldn't become accessible to search engines without proper protection.

The launch runbook includes SEO actions in the right order: a final reference crawl, the final URL export, deploying the redirects, checking the domain, robots, canonicals, sitemaps, tracking, and testing priority pages.

Every check has an owner. If BlackSwan works alongside an SEO agency, the split can vary: technical prep, business validation, mapping review, or monitoring. The responsibility table needs to reflect the actual contract.

Go/no-go criteria need to cover issues that would make the launch dangerous: major redirects missing, a global indexing block, incorrect canonicals, unusable tracking, or incomplete strategic templates. The exact list depends on the project.

The first few hours are for checking that everything works. The following days and weeks are for watching crawling, indexing, and traffic signals with enough distance to avoid jumping to conclusions.

Monitoring can combine crawls, Google Search Console, analytics, server logs where accessible, error tracking, and checks on priority pages. We compare the data against a dated baseline. Variations are analyzed by page type, market, device, or query when volume allows it.

BlackSwan can take charge of the technical setup and the agreed deliverables. The client often remains the owner of content decisions, access, business sign-off, and certain data. An SEO agency may be responsible for strategy, priorities, and final validation of the mapping.

There's no universal split. What matters is naming every owner before launch.

A migration like Claus Porto's, from Magento to Shopify with localized storefronts and market-specific catalogs, multiplies the URLs to carry over: languages, markets, product pages, and editorial content. This is exactly the kind of project the method on this page is built to protect.

A migration like Claus Porto's, from Magento to Shopify with localized storefronts and market-specific catalogs, multiplies the URLs to carry over: languages, markets, product pages, and editorial content. This is exactly the kind of project the method on this page is built to protect.