How to find broken links and fix them

To find broken links, check each important page with a link checker: it requests every link on the page and reports the ones that answer with an error, such as 404 Not Found or a 5xx server error. For a whole site, use a site crawler or Google Search Console's page indexing report. Then fix each link by pointing it at the right address, redirecting the old one, or removing it.

What is a broken link?

A broken link is a link whose target answers with an error instead of a page. The usual answers are:

404 Not Found: the page doesn't exist, often because it was moved or renamed.
410 Gone: the page was removed on purpose.
5xx errors: the server failed, either briefly or for good.
No answer at all: the domain expired, the DNS is wrong or the server is down.

A redirect is different. A 301, 302, 307 or 308 answer sends the visitor on to another address. The link still works, but it takes an extra step, and a chain of redirects adds more.

Do broken links hurt SEO?

Not in the way people often fear. Google's documentation says 404 errors are a normal part of the web and don't, on their own, hurt a site's indexing or rankings. Pages get removed, and search engines expect that.

Broken links still cost you something:

Visitors hit a dead end. Someone who clicks a link on your pricing page and lands on an error is less likely to keep going.
Crawlers waste their visit. Each dead link is a request that finds nothing. On a small site this barely matters; on a large one it can slow how quickly new pages get found.
Links to your own dead pages pass nothing. If other sites link to a page you deleted, that value is lost until you redirect the old address.

So the case for fixing them is mostly about people and housekeeping, not about a ranking penalty. Fix the ones on pages people actually read first.

How do I check one page in seconds?

For most makers, the pages that matter are a short list: the homepage, pricing, the signup flow, the docs landing page and any post you're about to share. Checking those one at a time is fast and catches the links people actually click.

A single-page link checker loads the page, collects every link on it and requests each one. The useful results are:

Links that answer with a 4xx or 5xx error.
Links that don't answer at all.
Links that redirect, so you can point them at the final address.
http:// links on an https:// page, which can trigger browser warnings or an extra redirect.
Links with no text, or vague text like "click here", which don't tell readers or search engines where they lead.

Run it before a launch and again after any redesign, since that's when links most often break.

How do I check a whole site?

Checking one page doesn't scale to hundreds. For a full site you have a few options:

Google Search Console is free and shows, in its page indexing report, pages Google tried to fetch and got "Not found (404)". It tells you which of your own URLs are dead, though not always which page links to them.
A desktop crawler such as Screaming Frog follows every link on your site and reports each status. Its free version is capped at 500 URLs, which is enough for many small products.
Hosted crawlers do the same in the cloud, usually with a free tier and a cap.

A good routine is to crawl the whole site every few months and check your key pages by hand whenever they change.

Your sitemap can help here too. It should list only pages that exist. If a sitemap lists URLs that now answer 404, search engines keep requesting them, so update the sitemap when you remove pages.

How do I fix a broken link?

Every broken link gets one of three fixes:

Update it. If the page moved, change the link to the new address. This is the best fix for links on your own site, because it removes the error and any redirect.
Redirect the old address. If other sites link to a page you moved or removed, set up a 301 redirect from the old URL to the closest matching page. Don't send everything to your homepage: search engines may treat a redirect to an unrelated page as a soft 404.
Remove it. If an external page is gone and there's no replacement, take the link out or replace it with a different source.

For outbound links to other sites, check whether the page moved before you delete the link. A search for the page's title often finds its new home.

What about redirect chains?

A redirect chain is when one redirect leads to another, such as http:// to https:// to www. to a new path. Each hop is an extra round trip, which slows the page down, and every hop is one more thing that can break.

Google's documentation says Googlebot follows up to 10 redirect hops before giving up. You don't want to get anywhere near that. Aim for one hop at most: link straight to the final address, and when you change URLs, update old redirects so they point directly to the newest page instead of stacking a new one on top.

How do I keep links from breaking again?

A few habits prevent most broken links:

Avoid changing URLs once a page is live. If you must, redirect the old one the same day.
Use relative links for pages on your own site, so a domain change doesn't break them.
When you remove a page, search your site for links to it and update them.
Re-check your key pages after every redesign, migration or CMS change.
Keep your sitemap in step with the pages that actually exist.

Broken links are a small problem that grows quietly. A few minutes before each launch keeps them from adding up. If you're preparing a launch, the pre-launch website checklist puts this alongside the other checks worth running.

Check it in a minute

Two free tools, no sign-up:

The broken link checker finds every link on one page, internal and external, and checks up to 50 of them, same-site links first and at most 10 on any other site. It reports broken (4xx, 5xx), unreachable and redirecting links, http:// links on an https:// page, and links with no text or text like "click here". For a whole site, pair it with a crawler or Search Console.
The sitemap checker finds your sitemap the way search engines do, through robots.txt or /sitemap.xml, checks that it's valid XML within Google's limits, reads its lastmod dates and flags URLs on another host, on plain http or written as relative paths.

Once your site is in shape, you can list your product on MakerRadar.