301 Redirects: What They Actually Do and Where Teams Get Them Wrong

Somebody on your team is about to delete a page. An old campaign landing page, or three blog posts that all say roughly the same thing and somebody finally decided to consolidate them. The question that comes next is always the same. Do we redirect it, and to where?

In more than 16 years of building content programs I have watched teams answer that question the same wrong way. Redirect it to the homepage, that way we do not lose anything. That instinct costs more than doing nothing, and the reason it costs more is the part nobody explains. Here is how redirects actually behave, which one of them quietly throws away everything the old page earned, and where the evidence sits inside Google Search Console.

What does a 301 redirect actually do

When a browser or crawler requests a URL, your server answers with a status code. A 200 means here is the page. A 404 means that page does not exist. A 301 means that page permanently lives somewhere else, and here is the address.

The consequence that matters is the second job. The old URL drops out of the index and the destination inherits what the old page had earned, which is why redirects matter for anything you have ever ranked with or built links to. Everything else in this post is a variation on one question. Does Google believe the statement your redirect is making?

The 308 status code does the same work with one technical difference. It preserves the original request method, so a POST stays a POST rather than being converted to a GET. For content pages that almost never matters. For form endpoints and APIs it does.

Do 301 redirects lose link equity

For most of the last decade SEO blogs repeated that every redirect bleeds off a percentage of link value, usually quoted around 10% or 15% per hop. Entire audit reports were built around chasing that number down.

Google guidance has been consistent for years now, and the redirects documentation states plainly that permanent redirects do not cause a loss in PageRank. If your only reason for cleaning up a redirect chain is a belief that it is quietly draining ranking power, you can stop worrying about that specific thing.

What you should worry about is more concrete. Every hop is another round trip. On a 150 millisecond connection, four chained redirects add roughly 600 milliseconds before the browser receives the first byte of the page the user actually wanted. That is before TLS, before HTML download, before render. Mobile users on slow networks experience that as a broken site, and they leave.

Which redirect actually costs you

This is the one nobody warns you about, because it never shows up as an error anywhere.

John Mueller has explained that a 301 makes sense when you have a genuine one to one replacement URL. When you do not, and you redirect an old page to something unrelated, Google can treat the redirect as a soft 404 and handle it like a 404. Nothing transfers. The signals the old page earned go nowhere.

This is exactly what happens when a team bulk redirects a few hundred retired URLs to the homepage. Search Console gets clean. The report shows Page with redirect instead of Not found 404, which feels like progress. But Google requested a page about a specific thing, got sent to a generic homepage with no relationship to the request, and drew the obvious conclusion.

You've converted a correct 404 into a soft 404, which is the weaker signal of the two. Do it at scale and you teach Google that your redirect rules are not reliable indicators of anything.

A 404 is not a penalty, and my guide to fixing not found 404 errors walks through why leaving one in place is often the correct call. A redirect is a statement that says this content moved here. If the content did not move, do not make the statement.

When should you use a 301 and when should you use something else

The row people skip is the canonical row. If both pages need to stay accessible to users, a redirect is the wrong tool because it removes one of them. Canonical tags consolidate ranking signals while leaving both URLs reachable, which is also the right instrument when two pages compete for the same query rather than duplicate each other. That distinction sits right next to the question of how many keywords one page can carry before it needs a sibling.

What do redirect chains really cost

A chain is what you get when the first redirect points at a URL that itself redirects. /old-post sends you to /old-post/, which sends you to /articles/old-post, which sends you to /articles/old-post/. Nobody designs this. It accumulates. A migration rule from two years ago, a trailing slash normalisation added at the CDN, a folder restructure nobody reconciled against the original rules.

On hop counts, the published guidance is genuinely inconsistent, and you should know that before somebody quotes a number at you in a meeting. Google site move documentation says Googlebot can follow up to 10 hops and recommends staying under three. Mueller has separately said Googlebot follows about five hops per crawl attempt and resumes on the next visit. At least one recent read of the current redirects documentation found no hop number stated at all.

Treat the number as unreliable and the direction as settled. One hop is the target. Anything past three deserves a look. Anything past five on a frequently crawled URL is a real risk, because a crawl attempt that does not resolve is a page that does not get indexed that day.

You can check any URL in about five seconds:

curl -sIL https://example.com/old-url | grep -i "^HTTP\|^location"

That prints every status code and every hop in order. More than one 301 before the 200 means you have a chain. Screaming Frog will map them across a whole site under Reports, Redirects, Redirect Chains, though you need to enable Always Follow Redirects first or it stops at your start folder. And if you want to see how crawlers are experiencing those hops rather than how you assume they are, server log analysis is the honest answer.

How long should you keep a 301 redirect in place

That timeframe exists because a site move happens URL by URL, and Googlebot has to crawl every affected URL at least once for the move to register. Until it does, nothing about that URL has moved as far as Google is concerned, which is the same mechanic behind getting your website indexed by Google and Bing in the first place.

The year is the search minimum, not the user minimum. Old links live on the internet indefinitely. A newsletter somebody sent in 2021 still points at your old URL, and removing the rule turns that into a 404 for a real person. Keep redirects indefinitely if the operational cost is low, and spend your effort updating your own internal links and any high volume external links so the redirect stops being load bearing.

How do you find redirect problems in Search Console

Page with redirect is a report rather than a warning, and Google is not asking you to fix it. Read the list anyway.

Redirected URLs still in your sitemap are you telling Google to crawl a page you have already moved. Internal links still pointing at old addresses spend crawl budget on a trip that did not need to happen. Chains do not get flagged as chains, but they surface the moment you spot check with curl. None of those will break your site. All of them are unforced.

The URL inspection tool turns any individual surprise in that list into a named diagnosis in about a minute, and it is the fastest way to confirm which URL Google actually treats as canonical after a move. If pages you redirected are sitting in crawled currently not indexed or discovered currently not indexed weeks later, the redirect is rarely the cause, but it is worth ruling out first.

One more boundary worth holding. Search Console covers Google. The index feeding ChatGPT search and Copilot is Bing, redirects have to resolve cleanly there too, and my complete Bing Webmaster Tools guide covers that parallel workflow. It matters more than it used to, because AI Overview citations regularly come from pages with no Google ranking at all. A redirect that resolves for Googlebot and stalls for Bingbot costs you visibility that shows up in neither traffic chart.

Frequently asked questions

Do 301 redirects hurt SEO?

No. A correctly implemented 301 to a genuinely equivalent page transfers ranking signals to the destination, and Google documentation states permanent redirects do not cause PageRank loss. The harm comes from redirecting to irrelevant destinations or building long chains, not from the redirect itself.

What is the difference between a 301 and a 302 redirect?

A 301 is permanent and tells Google to treat the destination as canonical. A 302 is temporary and signals that the original URL should stay indexed because the content is coming back. Using a 302 for a permanent move is a common misconfiguration that leaves the wrong URL in the index.

Should I redirect 404 pages to my homepage?

No. Google can classify an irrelevant redirect as a soft 404, which passes nothing and is a weaker signal than the 404 you started with. Redirect only pages that had real traffic or backlinks, and only to the closest genuine equivalent.

How many redirects is too many in a chain?

Aim for one. Published guidance varies between three and ten as the ceiling, and the numbers are inconsistent enough that you should not build a policy on the specific figure. Every extra hop adds latency and raises the chance a crawl attempt does not resolve.

When should I use a canonical tag instead of a 301 redirect?

Use a canonical when both pages need to stay reachable by users, because a redirect removes one of the two. Canonical tags consolidate ranking signals across near duplicate pages while leaving both URLs live, which is the right call for print versions, parameter URLs, and syndicated copies.

What is the difference between a 301 and a 308 redirect?

Both are permanent and both transfer signals. A 308 additionally preserves the original request method, so a POST stays a POST rather than being converted to a GET. For content pages the difference is irrelevant. For form endpoints and APIs the 308 is the correct choice.

How long does it take Google to process a 301 redirect?

It depends on crawl frequency, not on a fixed timer. Google has to crawl each affected URL at least once before the move registers for that URL, so heavily crawled pages transfer within days while deep pages can take weeks or longer.

Do 301 redirects pass link equity from backlinks?

Yes, when the destination is a genuine equivalent. Links pointing at the old URL count toward the new one. If the destination is unrelated and Google treats the redirect as a soft 404, those links pass nothing, which is why the destination choice matters more than the redirect type.

Can I remove a 301 redirect after a year?

For search purposes usually yes, because signals stay with the destination once the redirect has been crawled and in place for a year. For users no, because old links on other sites never expire. Keep the rule if the operational cost is low.

Do redirects affect how AI search engines cite my pages?

Yes, indirectly. ChatGPT search and Copilot retrieval sit on the Bing index, so a redirect that resolves cleanly for Googlebot but stalls for Bingbot removes the page from the citable pool. Test redirect resolution in Bing Webmaster Tools, not only in Search Console.