• Use 301 or 308 for permanent moves and 302 or 307 for genuinely temporary ones.
• Redirect to the closest equivalent page, one to one. Bulk redirects to the homepage are treated as soft 404s.
• Keep it to a single hop. Chains waste crawl resources and slow users down.
• Update internal links, canonicals, sitemaps and hreflang to point at the final URL.
• Keep redirects in place for at least a year, and ideally for good.

A redirect forwards a request for one URL to another. Done well, it carries users, links and ranking signals to the new location without anyone noticing. Done badly, it is one of the fastest ways to lose organic traffic.

Types of redirect

• 301 and 308 (permanent): tell search engines the move is final. The destination becomes the canonical and inherits the signals.
• 302 and 307 (temporary): tell search engines to keep the original URL indexed. Use these for short-term situations such as a temporary campaign or maintenance page.
• Meta refresh: an HTML-level redirect. An instant refresh is treated as permanent, a delayed one as temporary. Use only when server access is impossible.
• JavaScript redirects: work only if the page is rendered. They are slower to be picked up and are invisible to crawlers that do not run JavaScript.

Google confirmed years ago that 3xx redirects no longer lose PageRank simply for being redirects. What you can lose is relevance, if the destination does not match the original.

Implementation examples

Apache (.htaccess):

 Redirect 301 /old-page https://www.example.com/new-page 

Nginx:

 location = /old-page {
  return 301 https://www.example.com/new-page;
} 

Many sites now manage redirects at the CDN edge or in the CMS. The mechanism matters less than the result: a single server-side hop with the correct status code.

Best practice

• Map one to one. Each old URL should go to the page that best satisfies the same intent.
• Avoid chains. A to B to C should be collapsed to A to C. Googlebot follows up to ten hops, but every hop adds latency and risk.
• Avoid loops. A to B to A makes both URLs unreachable.
• Standardise the basics. http to https, non-www to www (or the reverse), trailing slash and lowercase rules should all resolve in one hop, not four.
• Fix the links. A redirect is a safety net. Internal links, canonicals, sitemaps and hreflang should point directly at the final URL.
• Preserve parameters when they matter so campaign tracking survives.

How long to keep redirects

Google's guidance is at least a year, which gives it time to see the redirect several times and transfer signals. In practice, keep them indefinitely where you can. External links to the old URL do not expire, and neither do users' bookmarks.

Common mistakes

• Redirecting everything to the homepage after a redesign.
• Using 302 for a permanent move.
• Redirecting to a URL that is noindexed, blocked or itself broken.
• Redirect rules that conflict, so the outcome depends on rule order.
• Letting an old domain expire, taking all its redirects with it.
• Redirecting users by IP address or language without giving crawlers a consistent route in.

How to test

• curl -IL shows each hop and its status code.
• Crawlers offer redirect chain and loop reports, and a list mode for testing a migration map before launch.
• Search Console's Page indexing report lists "Page with redirect" URLs, which is useful for spotting redirected URLs still sitting in sitemaps.

The GEO angle

AI models learn URLs from older training data and sometimes cite addresses that have since moved. Good redirects mean those citations still land. It is also worth checking your logs for 404s requested by AI user agents: assistants occasionally invent plausible-looking URLs on your domain, and redirecting the common ones to the right page recovers real visits.

Add your title here

This is a paragraph. Writing in paragraphs lets visitors find what they are looking for quickly and easily. Make sure the title suits the content of this text.