Broken links: what they are, how to find them and how to fix them

A broken link is a link that leads to a page that no longer exists or responds with an error, almost always a 404. To find them, combine the Page indexing report in Search Console, a link crawler and checks with curl. To fix them, correct the link, add a 301 redirect if there is an equivalent page, or remove it.

To check it on your own site: SEO audit

Broken links are a staple of SEO maintenance, and also one of its most exaggerated topics. This guide separates what Google actually says from what gets repeated without a source, and gives you a method for checking them without losing an afternoon.

What kinds of broken links are there?

A broken link is an <a href> whose destination returns an error or does not respond at all. Depending on where the link sits and where it points, there are three cases, and each is fixed in a different place:

Internal links matter most, because you control them completely and because they are how Google discovers your pages. Google uses links to find new pages to crawl and as a signal when working out how relevant they are1.

404, 410 and soft 404s: what does a broken link return?

A 404 says the server cannot find the page and does not say whether that is temporary or permanent; a 410 says the page has gone and probably will not come back. That is how the HTTP standard defines them, and it prefers 410 when the server knows the removal is permanent2.

For Google the difference does not matter. It treats all 4xx codes except 429 the same way: it does not index those URLs and drops the ones it had indexed3. Back in 2011 its blog already said it made no difference which of the two you returned4.

The case worth watching is the soft 404: a URL that does not exist but returns 200, or redirects to the home page. Google warns that such pages keep being crawled and waste crawl budget, and asks you to eliminate them5.

A word of caution before declaring a link broken: a 403 or 429 returned to an automated tool proves nothing. Many servers block or throttle crawlers that are not browsers, and the same URL opens fine in Chrome. Only 404 and 410 say the page does not exist; everything else needs a manual check.

Do broken links hurt SEO?

Not as a penalty, but they do break paths. Google explained on its Search Central blog that some URLs on your site returning 404 does not affect how your other pages perform in search results, and that 404s are a normal part of the web4. Search Console says the same: a 404 is not necessarily a problem if the page was removed without any replacement6.

What does have consequences is something else:

• A broken internal link to a page that now lives at another URL. The new page stops receiving that link, which is one of the ways Google finds and assesses it1.
• An inbound link that returns a 404. If another site links to you with a typo or to an old URL, Google suggests a 301 redirect to the correct URL to capture that traffic4.
• The visitor. Someone who clicks a menu link and lands on an error has to find another route or leave. We know of no study with a published methodology on how many leave, so we give no figure here.
• Crawling on large sites. Google recommends returning 404 or 410 for permanently removed pages, because a 404 is a strong signal not to crawl that URL again5. Its crawl budget guide is aimed at sites with a million pages, or ten thousand that change daily; on a small site this is not the main concern.

How do you find broken links?

Three sources, from the easiest to the most thorough. None of them sees everything on its own.

Google Search Console

In the Page indexing report, the "Not found (404)" reason lists the URLs on your domain that Google tried to crawl and got a 4046. To see where they are linked from, open each one in URL Inspection: under discovery it shows the "Referring page" Google may have used to find it7.

Its limit: it only sees URLs on your domain that Google knows about. Broken external links in your articles do not show up there.

A link crawler

A crawler goes through your site the way Googlebot would and checks every link, internal and external. The W3C Link Checker is free, runs in the browser and can check a whole site recursively, to the depth you choose8. Commercial desktop crawlers do the same with more filters and spreadsheet export.

curl, to check a list

If you already have the list of URLs (exported from a crawler or your CMS), this loop returns only the ones that fail or redirect:

# links.txt: one URL per line
while read -r url; do
  # -L follows redirects; -o /dev/null discards the body
  # %{http_code}: final status; %{num_redirects}: hops to get there
  result=$(curl -s -o /dev/null -L --max-time 15 \
    -w "%{http_code} %{num_redirects}" "$url")
  echo "$result $url"
done < links.txt | awk '$1 !~ /^2/ || $2 > 0'

# How to read the output:
# 404 0 / 410 0   -> broken: fix or remove the link
# 200 2           -> works, but after two redirects: link to the final URL
# 403 0 / 429 0   -> inconclusive: open it in a browser
# 000 0           -> no response (domain down or timed out): try again later

The second line matters more than it looks: an internal link that works after two redirects is not broken, but it makes every visit and every crawl take extra hops. The technical SEO guide covers redirect chains.

How do you fix each broken link?

There are three fixes: correct the link, redirect the URL or remove the link. Which one applies depends on whether the content still exists somewhere else.

One rule that saves trouble: redirect only if the destination answers what the person who clicked was looking for. A discontinued product page can go to the model that replaces it, not to the general category or the home page.

If the failure is your error page itself, because it returns 200, fix that first: until you do, no tool can tell your broken links from your working ones.

Links to other sites: why do they decay?

External links break on their own over time, even if you touch nothing. In October 2023, Pew Research Center checked a sample of almost a million pages collected by Common Crawl between 2013 and 2023. Of those that existed in 2013, 38% were no longer accessible10.

The same study checked the links on pages collected in 2023: 23% of news pages and 21% of government pages had at least one broken link, and 54% of the Wikipedia articles analysed had at least one dead link in their references10.

Three habits reduce the problem:

1. Link to stable sources. Official documentation and reference pages last longer than a press release or a campaign page.
2. Note when you consulted each source. If the link dies, you will know which version you cited and can look for an archived copy in the Internet Archive's Wayback Machine.
3. Check the external links on your most visited pages every few months, not just the internal ones.

How often should you check?

A small site that changes little can get by with a pass every quarter and another after each redesign, migration or removal of sections, which is when dozens of links break at once. For a shop or a publisher posting daily, monthly is better. To fold the check into regular tracking of several sites, see how to organise SEO monitoring. If it is your first time, do it as part of a full SEO audit.

What does SmoothSeen check?

Declaration of interest: this blog belongs to SmoothSeen, a web audit tool. When it analyses a URL, it checks the links on that page and flags those that lead to a 404 or a 410. It does not count links that return 403 or 429 as broken, because that is usually a block on automated tools rather than a missing page. You will find it alongside the other checks in what the search visibility analysis covers.

It does not crawl your whole site or connect to Search Console: for the full list of 404s Google knows about, the Page indexing report is still the source. The rest of what counts in Google is in the guide to what SEO is.

Frequently asked questions

Does a 404 error get you penalised by Google?

No. Google has explained that having URLs that return 404 does not affect how the other pages on your site perform in search results, and that 404s are normal when content is removed. What is worth checking is whether an important page returns a 404 by mistake, and whether your own links, especially in the menu, lead to pages that no longer exist.

Is it better to return a 404 or a 410?

For Google there is no practical difference: it treats both the same and stops indexing the URL. A 410 is more precise under the HTTP standard, because it says the page has been removed permanently, so use it if your server makes that easy. If you are unsure whether the removal is permanent, a 404 is the correct response.

Can I redirect every 404 to the home page?

It is not a good idea. Google considers redirecting every non-existent URL to the home page a soft 404, which makes your site harder to understand and index, and the visitor lands on a page without what they were looking for. Use a 301 only when there is an equivalent page; otherwise return a 404 with a helpful error page.

Do broken external links affect my rankings?

Google does not document any penalty for linking to a page that has disappeared. The effect is on readers, who cannot check the source, and on how credible the content looks. That is why it is worth checking them on the articles that get the most traffic and replacing each dead link with the document's new URL or another source that supports the same point.

What to do next

Open the Page indexing report in Search Console, go to the "Not found (404)" reason and use URL Inspection to see which pages link to each URL. Fix the links in your menu and templates first, as they repeat across the whole site. If you would like to see a page's broken links alongside the rest of its technical analysis, run a free SEO audit with SmoothSeen.