What a redirect is
A redirect is the server answering "that page lives elsewhere now" and forwarding the visitor —
and Googlebot — to the new address automatically. You see it when an http:// link lands you on
https://, or an old blog URL opens the renamed article.
The kinds that matter
- 301 — moved permanently. The standard for real moves: renamed URLs, domain changes, retired pages pointing to their replacement. It tells Google to transfer the old URL's standing to the new one and swap it in results.
- 302 — moved temporarily. For genuinely temporary situations (A/B tests, geo-routing). Google keeps the original URL indexed, expecting it back.
- Client-side redirects — a meta-refresh tag or JavaScript
window.location. They work, but slower and less reliably for crawlers than a proper server redirect; use the server kind when you control it.
Google has said 301s and 302s no longer lose PageRank in themselves. What still costs you is what redirects are attached to: extra requests for every visitor, and crawl effort for every hop.
Chains and loops
A chain is a redirect to a redirect: http://site.com → https://site.com →
https://www.site.com → the page. Each hop slows the visitor down, and
Googlebot follows a limited number of hops
(around ten) before giving up on the URL. A loop — A redirects to B, B back to A — makes the
page unreachable entirely.
The fix is the same for both: make every redirect point directly at the final destination, one hop total.
The habit that avoids all of this
Link, promote, and sitemap the final address — exact protocol, exact host, trailing slash and all. Redirects are a safety net for old links, not a routing layer for your own. If your own navigation, sitemap, or ad campaigns go through redirects, every visitor and every crawl pays the detour toll for no reason.