
A website redesign changes far more than what appears in a screenshot. Pages can acquire new addresses, navigation labels can change, old publications can be grouped differently and metadata may be generated by a different process. Visitors arriving through a saved link or a search result are still using the old structure. A successful migration should give them a meaningful way into the revised site instead of treating all existing addresses as disposable.
Separate the content migration from the visual decision as early as possible. Modern templates may make better use of information already published without requiring the address to change at all. When a route genuinely needs a new destination, establish the relationship between the old and new pages before release. A practical migration is built around an inventory and a map of decisions, not a blanket rule created after the revised navigation is complete.
Collect the existing route inventory
Begin by recording the addresses currently published. Include the homepage, service pages, category listings, detail articles and important older material. Use existing source records, sitemaps, crawl results and traffic data where these are available. They reveal different aspects of the site: a sitemap may be missing old addresses that still receive visits, while a crawl from the current navigation might overlook a useful article linked only from somewhere else.
For each route, record its title or purpose, publication state, current response and any known preferred address. Clarify whether it is unique content, an equivalent address for the same material or a genuine historical route with little remaining relevance. Where there is more than one version of a page, identify the version that presently contains useful content. This classification prevents the redirect list from treating every old string as an independent page.
Make a separate record of other links that the website relies on. Some article figures and downloadable materials may already have stable addresses that visitors use independently of their containing page. A migration that preserves every HTML entry but moves all essential files without replacements still loses important parts of the prior structure. The scope depends on the site. Include resources whose old addresses remain important, rather than trying to maintain an obsolete copy of every build file.
Distinguish moved content from removed content
A redirect should express a real relationship between destinations. Suppose an old publishing guide about image optimization has been expanded into a new tutorial on responsive images. If the new tutorial covers the original purpose, its address is a sensible target for the old guide. Visitors receive the material they expected in a revised form. The move should have a direct, permanent redirect where that reflects the intended structure.
Related reading: Testing a release beyond the happy path.
Now suppose a discontinued service description has no equivalent because the business no longer provides that service. Redirecting it to a popular article about an entirely different topic offers little practical benefit. The visitor asked for one thing and received another. For material genuinely removed with no useful replacement, a clear not-found response may be appropriate. A helpful error page can provide ordinary routes into the current collection without pretending that unrelated content is the new version of the old page.
Some groups of content may be combined. If several thin guides are meaningfully covered by a comprehensive new page, they can share that destination, provided the new material genuinely addresses each original subject. Review the titles and central questions before choosing the mapping. Similar vocabulary alone does not establish equivalence: an article on image compression and a service page for commissioning artwork refer to different needs even if both mention illustrations.
Map old addresses directly to final destinations
Create a route table with one row per old address and its intended disposition. A row may say that the route will stay, permanently move to another route or be removed. For a move, record the exact final destination and why it is appropriate. This simple record is useful during implementation and later review. When a page moves again, the old address can point directly to the newly preferred destination rather than participating in a growing chain of historical moves.

Choose durable conventions for the new paths, including trailing slashes and the preferred host. The consistency matters because visitors can discover an address through many different parts of the site. Navigation should use the final path directly, as should related-reading links. Avoid adding unnecessary redirects to fix inconsistent internal links that could be written correctly in the first place. The purpose of a migration redirect is to handle old entry points, not to become the normal way to navigate between current pages.
If historical pages exist in several variants, include the variants that genuinely need resolution. A former extension-based path and the current directory-style address may represent the same content. The route mapping should clearly distinguish them. Check that no rule sends a destination back to an old variant, and verify that broad pattern rules do not unintentionally capture a current page with its own content. A small number of precise mappings is often preferable to a rule that seems convenient but has uncertain scope.
Rebuild navigation around the final route map
A new category hierarchy can improve discovery without forcing every article to move into a new path. Menu labels, breadcrumbs and route conventions have related but distinct functions. A menu may adopt a clearer label while keeping the address of a useful service page. A breadcrumb can express the current content hierarchy even when an article retains its older stable route. Establish which changes serve the visitor and which are merely cosmetic preferences about the URL.
Create the new sitemap and canonical metadata from the published route inventory. Each normal indexable page should identify its own preferred address where that is appropriate. Retired addresses and redirect-only variants should not remain in the list of current pages. Confirm that the visible heading and descriptive metadata still correspond to the content being delivered. A migration can repair several structural inconsistencies at once if the templates derive those values from the same source record rather than independently maintained lists.
Set up a comparison release before switching traffic
Generate or deploy the revised site in a suitable test environment. Use a small set of representative routes to inspect the complete transition: a stable article, a moved service, a category, an obsolete page and a previously duplicated address. Request each expected destination and compare its status and canonical metadata with the route table. Check its navigation, visible title and internal links. A redirect to a perfectly valid current page can still be wrong if the destination no longer answers the original topic.
Also verify the asset relationships: a detail page should load its necessary fonts, styles and figures without referring to resources that exist only in the previous release environment. Inspect narrow and large layouts, especially for templates that changed during the migration. The older article content may be longer or otherwise different from the examples used to design the new templates. Include several such articles in the review rather than examining only newly written pages.
Keep a record for the next change
After deployment, verify representative old paths again against the live host. Check the final page rather than accepting the first redirect response as sufficient. Confirm the preferred destination, the response status of removed pages and the regenerated sitemap. Record any unexpected paths discovered after release and evaluate them using the same content-equivalence principle rather than automatically adding more broad redirects.
Preserve the route table with the website's publication source. It documents decisions that would otherwise be difficult to reconstruct when the original site is no longer available. A future maintainer can see that an old page was moved because its practical subject continued in a particular guide, or removed because no equivalent existed. Done carefully, a migration creates a more coherent current structure while maintaining useful links from the past. It does not need to make every historical route part of the normal navigation.
Explore the implementation area: Web development.


