SEO migration: before and after launch

Why a redesign can cost you traffic

Search engines know your pages by their URLs. If a redesign changes URLs without redirecting the old ones, those old URLs return an error: search results, links from other sites and your visitors' bookmarks all lead to a missing page.

URLs are not the only thing at stake. A redesign also rewrites titles, descriptions, canonical URLs and indexing directives. A noindex directive set on the staging site and left in place at launch removes pages from search results without breaking anything on screen.

Even a well-run migration moves rankings for a while: Google warns that visibility can fluctuate while it discovers the new URLs. Not every migration changes URLs, though. A change of hosting that keeps them identical needs no redirects, for instance a move to EU web hosting on the same domain.

What to settle before launch

Preparation comes down to three decisions, which Google describes in its guidance on site moves. They belong to the project itself, before the new site is live.

• An inventory of current URLs, including those no menu links to anymore but other sites still cite. A URL missed at this stage will get no redirect.
• A redirect map: each old URL leads to the new page that replaces it, in a single hop. Sending every old page to the homepage does not tell search engines where each piece of content went, and Google treats it as a missing page.
• How long to keep them: Google advises keeping a migration's redirects for at least a year, and longer for visitors who still use the old URLs.

A good map also avoids chains: the guide on redirect chains explains why a single hop matters.

What we find on the sites we audit

This percentage covers about 30 websites we audited between August 14, 2026 and September 24, 2026, the ones where this point could be checked. Many were audited because a defect showed up quickly, so the figure describes our audits, not websites in general.

Large numbers of broken internal links are rare on the sites we audit. Redirect chains are common: a chain forms when a change adds its own rule on top of earlier ones, and every migration is a chance for that to happen.

What the audit remeasures after launch

The audit crawls the live site in a browser, as a visitor would, and checks what a migration may have moved.

• Internal links: those that lead to a missing page or a server error. At 5% or more of the links followed, the report counts a defect. A page that asks visitors to log in is not counted as broken.
• Redirects on the site's entry URLs and canonical URLs: a single hop, permanent, with no loop.
• The tags the redesign rewrote: title, description, canonical URL, indexing directive, language versions, sitemap and robots.txt. The audit compares each page as the server sends it and as the browser displays it.
• A URL that does not exist: it should return a 404 error, not an ordinary page that search engines may index.
• A staging site left public and indexable, publishing a copy of your site at another address.

When the public web archive holds a record of it, the methodology note also flags a recent redesign, which helps read the results.

What you can catch up on, and what takes a project

Many oversights can be fixed after launch: a missing redirect on a page that used to get visits, a temporary redirect that should be permanent, a leftover noindex directive, internal links that still point to old URLs.

A migration done without an inventory takes more work. The redirect map then has to be rebuilt from what remains of the old site: your analytics, links from other sites, and copies kept by the public web archive.

What this check doesn't tell you

By Quentin Mathis, Z29K · updated September 28, 2026