Broken Link Example: Broken Internal Links
My rope down from the Moon ended a good way above the ground, and I could only observe the rest of the descent.
The short answer
- What it is
- A broken internal link points to a page on the same site that answers with a status of 400 or higher, or gives no response at all. Each such link counts on its own, even if many share a target.
- Why it matters
- Broken links strand visitors and waste crawl time. They also stop link value from reaching the pages you care about.
- How a crawler should report it
- Report one row per link with the source page, the target URL, the anchor text, and the status or connection error. Keep 429 and 999 responses out of this report, since those are signs of blocking and not of a dead page.
- Live examples
/lab/broken-links/(positive)/lab/broken-links/near-miss/(near miss)
Tracing a Dead Link Back to Its Source Page
Broken internal links are links on your own site that lead to a page that can't be served. A status of 404 means the server has no such page. A 410 says the page is gone for good. A 500 or 503 points to a server fault, and a timeout gives no status at all. The common causes are simple. A page was deleted or renamed. A slug was edited. A template was typed wrong, or a relative link resolves to the wrong folder.
A browser shows the visitor whatever error page the server sends. Most people leave at that point. Search engines drop a 404 or 410 page from their index over time, and they do not count an occasional one against the whole site. But each dead link wastes a crawl request. It also stops link value from flowing to the rest of your pages. Watch for soft errors too, where a missing page answers 200 with a message that says nothing is there.
Scale makes this worse. One bad link in a menu or footer can repeat across every page on the site. That is why reports count each link and not only each target. A server fault can also look like a broken link for a short time. Check a failing URL twice before you act on it, and use a GET request, since some servers mishandle HEAD.
To detect broken links, crawl the site from the home page, and record each link's source, target, and status. Sort by target to see which fix helps the most. To fix one, update the link to the right URL if the page has moved. Add a 301 redirect if outside sites still link to the old address. If the page is gone for good, remove the link. Then crawl again to confirm that the count has dropped.