A website migration may change the domain, URL paths, hosting platform, or several of these at once. Start by listing what will change and what must keep working, then plan how to test the new site and switch traffic to it. Include a way to return to the previous setup if the launch fails. No plan can guarantee zero downtime or unchanged search visibility.

Treat content, hosting, and data as related parts of the plan. Cutover is the point when the new system starts serving live traffic or accepting writes, and those switches need to happen in a tested order. A redirect can preserve access to a moved article, while database writes, mail delivery, scheduled jobs, and login sessions need their own continuity checks.

Inventory URLs and Dependencies

List the old site's pages, files, images, downloads, important API routes, and entry points. Use the content system, sitemap, traffic data, and server logs to find URLs that visitors or crawlers actually request. Include pages reached through external links or less common searches, not only the home page. For each URL, record its current status, content owner, and intended outcome after the move.

Inventory services that depend on the current domain or host: DNS records, TLS certificates, email routing, payment callbacks, OAuth return URLs, webhooks, analytics, content delivery network (CDN) settings, background jobs, and administrative access. A public page can appear healthy while a callback still points to the old host. Identify the team responsible for each dependency and the test used to check that the service works on the new site.

Record a baseline before changing anything: page availability, completed purchases or registrations, error rates, crawl activity, and data counts. Back up the current site and restore that backup in a separate test environment. Confirm that the restored application can read its data and start correctly, and measure how long recovery takes.

Map Old URLs to New Destinations

If URLs will change, create an old-to-new URL mapping. A hosting move that preserves every public URL does not need redirects merely because the server changes. Match each old page to the closest equivalent new page, preserving the visitor's intent. If a guide has moved, redirect to that guide rather than to the new home page. If content has been removed without a relevant replacement, return a suitable 404 or 410 response instead of redirecting visitors to unrelated content. Include media and downloadable files where they have independent URLs or links.

Mark rules that can be expressed as a pattern and exceptions that need explicit entries. Test ordinary paths, encoded characters, query parameters, and exceptions to each rule. Check that the rule does not also match its own destination. Prevent redirect loops, extra hops, and accidental HTTP downgrades. Use permanent server redirects for URLs that have moved permanently, and check how the chosen status code handles forms or API requests as well as ordinary page visits.

On the new site, update internal links, canonical annotations pointing to the preferred new URLs, sitemap entries, and language annotations where used. Keep the destination accessible and indexable when it is meant to appear in search. A staging noindex rule or a restrictive robots.txt file must not be carried into production by accident.

Stage the New Site and Verify It

Build the new environment before directing public traffic to it. Check page content, templates, media, forms, search, API responses, keyboard access to forms and navigation, and mobile layouts. Validate TLS certificates for every hostname visitors may request, including redirecting names. Test with realistic but non-sensitive data and document expected versus observed results.

For a site with changing data, decide which system may accept new orders, edits, and uploads at each stage. One option is to stop writes on the old system, copy the final changes, validate the new data, and then open the new system for writes. Replication can reduce the amount left to copy, but the plan must account for any changes that have not yet reached the destination.

Keep database schema changes compatible with the application version that will use them. Check uploads, queued work, and session storage as well as database tables. Unless the application is designed to coordinate jobs across both environments, stop scheduled jobs on the old system before starting their replacements so the same work is not performed twice.

Prepare the redirect rules and DNS changes without assuming DNS updates are instantaneous. Resolver caches can continue sending some visitors to the old endpoint after the authoritative records change. During that overlap, a changed URL can redirect to its new address. If the public URL stays the same, keep the old server serving consistent content or forwarding requests internally to the new system; redirecting a URL to itself only creates a loop. Record current and planned DNS records, their TTL values, and how you will check requests reaching each server.

Cut Over in a Controlled Window

Choose a launch window with staffing and a rollback decision maker. Lower traffic may reduce visitor impact, but the site still needs active monitoring. Freeze unrelated feature and design changes where practical so an error has fewer possible causes. Follow the sequence rehearsed for the site: complete the final data copy or replication checks, confirm where writes are allowed, then switch routing and enable any required redirects. Change DNS only if the move requires it, and verify the responses that visitors actually receive.

Run a short checklist immediately after cutover: home page, top landing pages, less frequently visited pages, images and downloads, forms, login, checkout or other critical transactions, API callbacks, and email or background services. Where URLs changed, confirm that the old addresses lead to the mapped pages and that internal links use the final URLs. Also check that new pages return the intended status codes, rather than showing error content with a success response. Test from more than one resolver or network when DNS is part of the move.

Search rankings can fluctuate while changed URLs are recrawled and reindexed. Track search visibility alongside page errors, response times, transactions, and support reports. A working home page does not show whether all migrated pages and services are working.

Monitor the Move and Keep Rollback Available

Watch old and new traffic, redirect errors, server logs, crawl reports, indexing signals, and application health. Group errors by path or file type to spot failures affecting a whole section. Compare against the pre-move baseline while allowing for expected shifts in URL reporting. Keep the old domain and redirects available long enough for visitors and crawlers to update their references.

Define rollback triggers before launch: for example, a critical transaction failure, widespread 5xx responses, or data divergence that cannot be repaired safely in place. State exactly which DNS, proxy, application, and data changes can be reversed and in what order. If the new system has accepted orders or edits, preserve and reconcile those changes before returning to the old database. Reverting DNS cannot copy that data back or clear permanent redirects already cached by clients. Practice the recovery path in staging and keep a current snapshot of the prior configuration.

Once the new site is stable, update external profile and campaign links where possible. Before canceling old hosting, check that no mail service, job, file store, or incoming traffic still depends on it, and keep any required redirects running elsewhere. Save the final URL map, configuration, test results, and remaining tasks so a later problem can be traced to the actual changes.