A page full of dead links reads to Google as a site nobody's maintaining anymore. Drop in any URL and see exactly which links are actually broken, which just redirect, and which ones return 200 OK while quietly serving a login wall instead of the page you linked to.
Google reads a page's link health as a proxy for how well the whole site is maintained. A handful of dead links doesn't just annoy the occasional visitor who clicks one; it's read as a small but real signal that nobody's keeping the place tidy.
Three things happen every time a link on your site breaks, and none of them are obvious from looking at the page itself:
A browser extension that only checks what's visible on screen misses everything in the footer and nav that a crawler still follows. This reads the full HTML source instead, the same way a search engine would.
<a href> in the page gets parsed, including the ones tucked into navigation, sidebars, and footers that a human rarely clicks but a crawler always follows.Not every broken link gets the same treatment. What you do depends on where it lives, what error came back, and whether you control the destination at all. Worth working through in roughly this order.
Fully within your control, so fix these first. Restore the page if it still has value. If the content moved, set up a permanent 301 redirect to whichever existing page is the closest topical match, not a blanket redirect to the homepage. Google treats homepage-catch-all redirects as a soft 404 and discards them outright.
A 410 means the resource was deliberately removed for good. If your own pages return this, strip them out of your XML sitemap, navigation, and any content still linking to them. Google drops 410s from the index and stops re-crawling them faster than 404s, so a live link pointing at one is just spending a crawl visit on a confirmed dead end.
A link going A to B to C should just point at C. Each extra hop in between costs a fraction of the authority passed through it and adds real latency on a large site. The Redirect Checker traces the full path so you know exactly where C actually is before updating the source link.
You can't fix someone else's site, but you can choose what you link to. Look for an updated URL on the same domain, an archived copy via the Wayback Machine, or a different source entirely. If nothing fits, removing the link beats leaving a dead citation sitting in your content.
On an external site, a 5xx might just be a brief outage or rate limiting, worth rechecking in a day before calling it broken. On your own domain, treat it as urgent immediately: Googlebot stops crawling a URL after repeated server errors, and it can drop out of the index before anyone notices the pattern.
Some links resolve fine technically while serving a login wall, a paywall, or a "this content has moved" placeholder instead of the original page. The status code looks healthy, so this one needs an actual look at what's rendering, not just a glance at the response code.
This scan is a snapshot. On a site that's actually growing, links keep breaking after it: pages get deleted, URLs change, third-party sites go offline. Re-running a manual check every week doesn't scale past a certain point.
✓ Free plan available · ✓ No credit card required · ✓ Set up in 2 minutes