Redirects are among the most consequential technical SEO decisions a site can make. A site migration executed with correct redirects preserves organic traffic. The same migration executed with redirect chains, wrong redirect types, or missing redirects can cause significant ranking losses that take months to recover from. Understanding redirect mechanics and when each type is appropriate is foundational technical SEO knowledge.
How Redirects Affect PageRank
Google transfers PageRank through redirects, but not identically across redirect types.
301 (permanent redirect): Signals the destination URL is the permanent home of this content. Google consolidates ranking signals — including backlink equity — to the destination URL. This is the standard redirect for permanent URL changes, site migrations, and consolidating duplicate content.
302 (temporary redirect): Signals the redirect is temporary and the original URL may be used again. Google keeps the original URL in its index rather than consolidating signals to the destination. Use cases: A/B tests, maintenance pages, geolocation redirects where the original URL is the canonical.
307 (temporary redirect, HTTP/1.1): Functionally similar to 302 for SEO purposes. Preferred over 302 for HTTPS-specific contexts.
308 (permanent redirect, HTTP/1.1): Functionally similar to 301 for SEO purposes, but preserves the HTTP method (POST requests stay POST). Relevant for applications; rarely needed for pure content redirects.
Meta refresh: A client-side redirect defined in HTML. Slower (requires the page to partially load), less clean than server-side redirects, and introduces latency. Avoid for SEO-critical redirects.
JavaScript redirects: Client-side redirects via window.location. Treated like page content — Googlebot must execute JavaScript to follow them, which introduces the same delays as other JavaScript-dependent content. Not equivalent to server-side 301s for SEO purposes.
Redirect Chains: The Hidden Crawl Problem
A redirect chain occurs when URL A redirects to URL B, which redirects to URL C. Each hop in the chain:
- Adds HTTP request latency
- Risks losing PageRank (each redirect dilutes equity slightly)
- Wastes crawl budget (Googlebot follows each hop separately)
- Can cause chain-following limits (browsers and crawlers may not follow chains longer than 5–10 hops)
Common cause: incremental site migrations. A URL changed once (A → B), then changed again (B → C), without updating the original redirect. After three migrations, chains can be 4–5 hops long.
Detecting chains: Crawl tools (Screaming Frog, Sitebulb) report redirect chains with hop counts. Run a full site crawl and filter for 3xx chains longer than 1 hop. Export and fix by updating each originating redirect to point directly to the final destination URL.
Checking inbound backlinks: External links pointing to intermediate URLs in a chain pass equity less efficiently than links pointing to the final URL. After collapsing chains, update high-value backlinks if possible to point directly to the final destination.
HTTPS Migration: The Most Common Redirect Scenario
The HTTP → HTTPS migration is the most common bulk redirect scenario. Best practices:
- Implement server-level redirects (not meta refresh) for all HTTP URLs to HTTPS equivalents
- Use 301s, not 302s — this is a permanent change
- Update internal links throughout the site to use HTTPS directly (reduces redirect hops)
- Update canonical tags, sitemaps, and structured data to use HTTPS URLs
- Verify Google Search Console has the HTTPS property verified and set as the primary property
A correctly executed HTTPS migration preserves organic traffic almost entirely within the crawl cycle. Poorly executed migrations (302s, chains, missing internal link updates) can cause weeks of ranking disruption.
Site Migrations and URL Changes
Larger site migrations — platform changes, domain changes, URL structure changes — follow the same principles at scale:
Prepare the redirect map before launch. Export all indexed URLs (from GSC, sitemaps, crawl tools) and map each to its destination URL. Every indexed URL should have a redirect — not just the homepage.
Use exact URL matching, not pattern-only. Pattern-based redirects (/old-section/* → /new-section/*) can miss edge cases. Verify the map with a crawl after launch.
Redirect old canonical URLs. If the old site had canonical tags pointing to specific URLs, redirect those canonical URLs even if the content lives at different URLs post-migration.
Monitor GSC immediately post-migration. Coverage errors, crawl anomalies, and indexation drops in the days after launch indicate redirect problems. Early detection shortens recovery time.
Don’t redirect everything to the homepage. A common mistake: redirecting all old URLs to the homepage when specific content was removed. This is a soft 404 pattern — Googlebot sees a 301, follows it to the homepage, determines the page content doesn’t match the original URL, and treats it as a missing page.
Redirect Auditing at Scale
For large sites, maintaining redirect hygiene requires systematic processes:
Quarterly redirect chain audits: Run a crawl, export all redirect chains, collapse any chains longer than 1 hop. Schedule this as a regular maintenance task.
Backlink redirect monitoring: Periodically check high-value backlinks (from Ahrefs/Majestic/GSC) to verify they point to live pages, not redirect chains or 404s.
Post-deployment redirect checks: After any site change that involves URL changes, run a targeted crawl of affected URL patterns to verify redirects are in place and correct before Google recrawls.
Crawl log analysis: Server log analysis shows which redirects Googlebot is actually following and at what frequency. Redirects being followed repeatedly by Googlebot (rather than just once before the destination is cached) indicate the redirect isn’t being processed as permanent.
When Not to Redirect
Not every old URL needs a redirect. If a page was removed and there is no relevant destination, a 410 (Gone) response is semantically cleaner than a redirect to an unrelated page. 410 tells Googlebot the page is intentionally gone and can be deindexed quickly — faster than a redirect to an unrelated page that Googlebot will eventually recognize as a soft 404.
Redirects are expensive to maintain and can accumulate as technical debt. Only implement redirects where there’s a meaningful destination for the content — either the same content at a new URL, or genuinely related content that serves the same search intent.