A 302 redirect passes PageRank and link equity. Google’s John Mueller called the opposite a myth in February 2016, and Gary Illyes said that July that 30x redirects don’t lose PageRank. What a 302 changes is which of the two URLs Google keeps. Google puts the redirecting URL and its target in one group, reads a 302 as a hint to prefer the source and a 301 as a hint to prefer the target, and the links pointing at either count for whichever URL it shows. The 302 that costs something is the one that keeps an old URL in the results after a move that was never coming back, and most of those were never chosen: a framework sent a 302 because nobody specified a code.
Below are the statements the “302 passes PageRank” answer rests on and which of them are documentation, how the URL choice works, where accidental 302s come from, and how to check in Search Console which URL Google picked.
What a 302 redirect is, and why SEO guides warn against it
A 302 redirect is the HTTP status code “302 Found”: a temporary redirect that sends the browser to another URL and tells it the original address is coming back. A 301 redirect, “Moved Permanently”, says the move is for good. For years SEO advice treated the two as opposites for links: a 301 was said to pass link juice from the old page’s backlinks to the new one, and a 302 was said to pass none, leaving the new page without the authority the old one had earned. That belief is where the question “do 302 redirects hurt SEO” comes from, and the first half of it has been out of date since 2016.
What Google has said about 302 redirects and PageRank
The case for 302 redirects passing link equity is a series of statements by Google staff since 2016, plus one line in the redirects documentation. They agree with each other, and it is still worth knowing which is which.
| Date | Who | What was said | Where |
|---|---|---|---|
| 3 February 2016 | John Mueller | “it’s incorrect that 302 redirects wouldn’t pass PageRank. That’s a myth.” | Google+ thread, reported by Search Engine Roundtable |
| 23 March 2016 | John Mueller | “There’s still confusion on 302/301 from the olden days; 302s are fine” | |
| 26 July 2016 | Gary Illyes | “30x redirects don’t lose PageRank anymore.” | |
| 8 April 2021 | John Mueller | A 302 is “source-preferred”, a 301 “destination-preferred”; on when a 302 gets treated as a 301, “There’s no fixed cut-off time for that.” | Twitter, reported by Search Engine Roundtable |
| 31 August 2023 | John Mueller | “The good news is it doesn’t matter. Use the technically correct redirect type.” | Google Search Central video |
| Current | Google Search Central | 301 and 308 are a strong signal that the target is canonical; 302, 303 and 307 are a weak one | Redirects and Google Search documentation |
Google’s site-move documentation says that 301 and other permanent redirects don’t cause a loss in PageRank. It has no equivalent sentence about temporary ones, so on 302s the record is what Google’s people have said, consistently for ten years, and the documentation’s description of a 302 as a weak canonical signal. I treat that as settled for PageRank. Geoff Kenyon, whose 2013 tests are still the only published experiment on 302s, updated his post in June 2026 and still calls the equivalence not fully settled in practice, and what he saw in practice is the URL choice described in the next section.
The percentages that still circulate, a 301 passing 90 or 99 percent of link equity and a chain losing 15 percent a hop, belong to the period before 2016. None of the pages repeating them cites a source from after it.
How Google handles a 302 redirect: one group, two URLs
Mueller’s 2021 explanation is the most useful model of what happens to link equity through a 302. Google puts the redirecting URL and the target into one group and runs canonicalization on it: it picks one URL of the pair to index and show, and the ranking signals of both, backlinks included, are consolidated on that one. The redirect type is an input into that choice. A 302 says the source is preferred, because it is coming back; a 301 says the target is. Mueller added that rankings come out much the same whichever URL Google shows, which is the part that gets summarized as “302s pass PageRank.”
The redirect type isn’t the only input. In the same thread Mueller said that if all internal and external links point to the destination, Google should probably pick the destination too. The other canonical signals vote the same way: which URL is in the sitemap, what the target’s rel="canonical" says, which URL the site’s own navigation uses. That is how a 302 left in place for a year ends up behaving like a 301. The site keeps voting for the target until the votes outweigh the redirect’s hint, and Mueller’s answer on the timing was that there is no fixed cut-off. A guide that names a number of weeks is guessing.
Where a 302 redirect costs you, if not in PageRank
The losses blamed on 302s are real, and they come from the URL choice. Four of them show up in audits.
- The old URL stays in the results. In Kenyon’s test of a 302 to an unrelated site, the original URL kept ranking, under its old title and with the other site’s meta description. A SERP after a careless 302 often looks like that: the source’s address on the target’s content.
- A migration that doesn’t finish. A site move done with 302s tells Google the old URLs are coming back, so the new ones replace them late, and only where the other signals overrule the hint. The dip that follows gets read as lost link equity while the links are intact and Google is still showing the old addresses.
- Reporting that splits. Search Console assigns performance data to the canonical URL, so a page whose 302 source is still the canonical reports its clicks under the old address. Compare before and after on the new URL and you see a drop that didn’t happen.
- Signals that contradict each other. A 302 to a target whose canonical tag points back at the source tells Google two opposite things, and Google settles it by choosing, differently from one URL to the next.
A redirect to a page that doesn’t replace the source is a separate failure, and the status code doesn’t fix it: Google is likely to treat that redirect as a soft 404 whether it is a 301 or a 302. The five ways a redirect gets ignored are in our 301 redirect SEO penalty guide.
Where accidental 302 redirects come from
Most 302s I find in audits weren’t chosen. The redirect was written without a status code, and on most server platforms and web frameworks the default is temporary.
| Platform | Written without a code | Sends | Permanent version |
|---|---|---|---|
| WordPress | wp_redirect( $url ) | 302 | wp_safe_redirect( $url, 301 ) |
| PHP | header("Location: /new") | 302 | header("Location: /new", true, 301) |
| Apache (mod_alias) | Redirect /old /new | 302 | Redirect 301 /old /new |
| nginx | rewrite ^/old$ /new redirect; | 302 | return 301 /new; or the permanent flag |
| Express | res.redirect('/new') | 302 | res.redirect(301, '/new') |
| Django | redirect('/new') | 302 | redirect('/new', permanent=True) |
| Ruby on Rails | redirect_to "/new" | 302 | redirect_to "/new", status: :moved_permanently |
| Laravel | redirect('/new') | 302 | redirect('/new', 301) or Route::permanentRedirect |
| Next.js (App Router) | redirect('/new') | 307 (303 inside a Server Action) | permanentRedirect('/new'), or permanent: true in next.config.js, both 308 |
| Shopify | URL redirects in the admin | 301 | already permanent |
Redirect plugins and CDN rules usually ask which code to use. Hand-written redirects, in application code or in an .htaccess file, are where the temporary default slips in, in login flows, in a quick fix during a launch, in a middleware someone wrote to move one section. Two searches find them: the codebase for the calls in the middle column, and a crawl for the result. In Screaming Frog, Response Codes → Redirection (3xx) with the Status Code column visible lists every 302 and 307 on the site.
Why developers reach for a 302 first
The habit has a reason, and the reason is the browser. Under the HTTP specification, RFC 9110, a 301 and a 308 are cacheable by default: a browser that has seen one can go straight to the target on the next visit without asking the server. A 302 and a 307 aren’t cached unless the response headers say so. A wrong 301 sticks in every browser that saw it, and a wrong 302 disappears with the next request, so developers test with 302 and intend to switch. The switch is the step that gets skipped, and a redirect that was temporary for testing stays temporary for years.
When a 302 is the right redirect
Whenever the source URL is coming back. Google’s documentation on A/B testing, last updated in December 2025, says to “use a 302 (temporary) redirect, not a 301 (permanent) redirect” for a test that sends users from the original URL to a variation, so that search engines keep the original in the index. The same logic covers a login wall, a product temporarily out of stock pointing at its category, a maintenance page and a short promotion.
Seasonal pages are the edge case. Google’s advice since 2020 is one recurring URL per event, /sale/black-friday rather than a new address each year, so that its links build up. The safest state between events is the page staying live with the next date on it. A team that insists on redirecting it off-season should use a 302, because the source is exactly what Google is supposed to keep. I haven’t measured whether a page treated that way returns faster than one that stayed live, and I wouldn’t promise it.
Country and language redirects are the case where 302 is the right code on a technique that causes its own problem. Redirecting by IP also redirects Googlebot, which crawls mostly from US addresses, so the other locales go uncrawled whatever status code sends it there.
302 vs 307 vs 301 vs 308
| Code | What Google reads | Request method | Cached by browsers by default | Use when |
|---|---|---|---|---|
| 301 | Strong signal that the target is canonical | Browsers may turn POST into GET | Yes | A page has moved for good |
| 308 | Strong signal that the target is canonical | Preserved | Yes | A permanent move where forms or an API post to the old URL |
| 302 | Weak signal that the target is canonical | Browsers may turn POST into GET | No | The source URL is coming back |
| 307 | Weak signal that the target is canonical | Preserved | No | Temporary, and the request method has to survive |
One 307 never leaves the browser. On a site with HSTS, Chrome upgrades an http link to https internally and shows it in developer tools as a 307 Internal Redirect. Googlebot doesn’t receive that one; what it gets for the http URL is whatever your server sends, which should be a 301.
How to check which URL Google picked for a 302
- List the temporary redirects. Crawl the site and filter Redirection (3xx) to 302 and 307. For each one, write down whether the source is coming back. If it isn’t, the redirect should already be a 301, and the rest of the check tells you what Google did in the meantime.
- Inspect the source URL. In URL Inspection, the Google-selected canonical field answers the question directly. If it shows the source, Google is keeping the old URL. If it shows the target, the 302 is being treated as permanent.
- Inspect the target URL. Check that it is indexed in its own right, and that its canonical tag, its place in the sitemap and the internal links all point at itself. A target whose own signals point back at the source is the contradiction from the list above.
- Read the Links report three to six weeks after a change. The referring domains that linked to the source should start appearing against the URL Google chose. Glenn Gabe showed in 2024 that the report credits links to whichever URL Google has canonicalized, which makes it a usable trace of the choice.
Changing a 302 to a 301 after this doesn’t recover link equity, because none was lost. It changes the hint, so the site’s other signals should change with it: internal links, the sitemap and the target’s canonical, all on the target, on the same release.
302 redirects in chains and on external links
A 302 in the middle of a redirect chain of 301s puts a weak signal between two strong ones, and the chain as a whole stops saying “permanent.” Googlebot follows up to ten hops, so the chain rarely stops the crawl; what suffers is the hint and the crawl spent getting there. Since Google rewrote its crawl-budget documentation in July 2026, that crawl is one allowance shared by all of Google’s crawlers. The AI crawlers spend more on hops than Googlebot does: in Vercel’s crawler data from December 2024, ChatGPT’s crawler spent 14.36% of its fetches following redirects, against 1.49% for Googlebot. Every hop removed from a chain is a fetch spent on a page instead.
External links are the other place 302s hide. Affiliate trackers, link shorteners and ad redirects often sit between someone else’s page and yours as a 302. They work as links through a redirect like any other, but they never give the target a permanent signal, and the code lives on a server you don’t control. The status code also doesn’t change what a link is: an expired domain or a network of sites pointed at yours with a 302 is the same link scheme under Google’s spam policies as one pointed with a 301.
302 redirect FAQ
Do 302 redirects hurt SEO?
Not through lost link equity. A temporary redirect used for something temporary is what Google expects and costs nothing. A 302 hurts SEO when the move was permanent: the old URL stays in the index, the new one takes longer to rank under its own address, and reports split between the two.
How long before Google treats a 302 as a 301?
There is no fixed time, and Mueller said so in 2021. What moves it is the rest of the signals: internal links, the sitemap, the target’s canonical tag and external links all pointing at the target. On a site where all of those still point at the source, a 302 can keep the source in the index indefinitely.
Does a 307 pass PageRank the same way as a 302?
Google’s redirects documentation groups 302, 303 and 307 as the same weak canonical signal. The difference between them is HTTP: a 307 guarantees the browser repeats the same request method, which matters for forms and APIs and has no effect on rankings.
Should I change old 302 redirects to 301s?
Where the move is permanent, yes, because the hint should match the intent. Expect the change to move which URL Google shows over the following weeks. Leave the 302s where the source is coming back: tests, login walls, temporary stock pages.
What I don’t know is how long, on a typical site, Google takes to switch a long-standing 302 over to the target. Google hasn’t published a number, I haven’t seen a dataset that measures it across sites, and the answer depends on how many of the site’s own signals still point at the source. To list every 302 and 307 on a site with its full chain, our redirect checker takes up to 500 URLs at a time.