Broken External Links Example
I climbed the Turkish bean to the Moon, and the path back down has been unresolved ever since.
The short answer
- What it is
- A broken external link points to another site and gets a status of 400 or higher, or no response at all. DNS failures, timeouts, and TLS errors all count as no response.
- Why it matters
- A dead outbound link lets readers down and makes a page look neglected. It is usually a small quality problem, but it grows as a site ages.
- How a crawler should report it
- Report one row per link with the source page, the external URL, and the status or error name, such as DNS failure or timeout. Keep 429 and 999 responses in a separate blocked group, since the page may be fine.
- Live examples
/lab/broken-external-links/(positive)/lab/broken-external-links/near-miss/(near miss)
When Outbound Links Stop Leading Anywhere
Broken external links happen because other sites change without telling you. This slow decay is often called link rot. A company goes out of business and its domain lapses. A publisher moves to a new CMS and its old paths vanish. A site drops its HTTP version, or its certificate expires. A link can also be wrong from the first day, due to a typo or a pasted tracking URL that has since expired.
A visitor who clicks a dead link sees an error page, or a browser warning if the host has failed. The trust you built can fade right there. Search engines don't pass a site-wide penalty for an outbound link that has died. Still, a page full of them signals poor upkeep. There is also a hidden risk. An expired domain can be bought by a stranger, and your old link then points to whatever they publish.
False alarms are common with external checks. Some servers refuse HEAD requests but answer GET normally. Others block unknown bots with a 403. So a result is only a lead until you confirm it. A 429 or 999 response means that the site is limiting you, which says nothing about the page. A short outage can also look like a dead link, so retry on another day.
To detect broken external links, collect all outbound URLs in a crawl and test each one with a browser-like GET request, at a slow rate per host. Retry the failures before you report them. To fix a confirmed one, find the new address on the same site and update the link. If the page is gone, replace it with a current source that covers the same point. You can also link to an archived copy. If nothing fits, remove the link and adjust the sentence.