External Links Returning 4XX: How to Triage and Fix Them

What this check flags

This audit item lists every outgoing link from your site to an external URL that responded with a 4XX status code during the crawl. A 404 (Not Found) or 410 (Gone) means the page no longer exists at that address. A 403 (Forbidden) means the server refused the request, sometimes indicating bot blocking rather than a truly dead page. Either way, the crawler followed a link you published and hit a wall. This is different from an external link that redirects to a broken destination, covered separately in our guide to external redirect broken errors.

Honest severity

Broken outbound links are not a direct ranking factor. Google does not penalize you for linking to pages that have since died, because link rot is universal and search engines know it. If you expect fixing these to move rankings next week, recalibrate.

The real cost is user experience and perceived quality. A reader who clicks your "according to this study" link and lands on a 404 learns two things: they cannot verify your claim, and nobody has touched this article in a long time. That second impression is the dangerous one. Dead references visibly age your content, and citations that no longer resolve quietly undermine the E-E-A-T perception you are working to build with both readers and quality raters.

Why external links rot

Link rot is not an edge case, it is the default trajectory of the web. A 2024 Pew Research Center study found that 38 percent of webpages that existed in 2013 were no longer accessible a decade later. Earlier Harvard Law research found that about half the URLs cited in U.S. Supreme Court opinions no longer led to the originally referenced material. If even legal citations rot at that rate, your blog references have no chance of staying pristine on their own.

The mechanics are mundane. Companies fold and domains expire. Sites migrate platforms and change slug structures without redirecting old URLs. Articles get unpublished, merged, or moved behind paywalls. None of this is your fault, but the broken link lives on your page, so the cleanup is your job.

Triage at scale

If your crawl surfaced hundreds of these, do not fix them in alphabetical order. Start with pages that actually receive traffic and impressions, because a broken link nobody sees costs you nothing today. Then prioritize in-content editorial references over boilerplate: a dead citation in your most-read guide matters far more than a stale link in a footer widget or author bio. Finally, a broken link that anchors a key claim deserves attention before a passing "see also" mention.

Fix options, in order of preference

For each link worth fixing, work down this ladder:

1. Find where the content moved. Often the page still exists at a new URL. Search the destination site for the title, or search the web for an exact phrase from your anchor text. If you find it, update the href and you are done.

2. Swap in a fresh source. If the original is truly gone, a newer source making the same point is often an upgrade, and the page reads as maintained rather than merely repaired.

3. Link an archive.org snapshot. When the exact original matters, such as a study you quote directly, the Wayback Machine usually has a copy. Link the snapshot URL so readers can still verify the claim.

4. Unlink. If the reference no longer earns its place, remove the anchor tag and keep or trim the surrounding text. A plain mention is better than a dead link.

Beware soft 404s

Crawlers can only flag what returns a 4XX status. Plenty of dead pages return a 200 with a "content no longer available" message, a redirect to the homepage, or an empty template. These soft 404s are invisible to status-code checks, so for important citations, click through and confirm the page still says what you cited it for, not just that it loads.

How to diagnose

Before editing anything, verify the error is real. Crawlers get blocked by firewalls and bot protection, so a 403 in your audit may be a healthy page. Test with a browser-like request:

curl -sI -A "Mozilla/5.0" -L "https://example.com/dead-page" | head -5

# Check for an archived copy
curl -s "http://archive.org/wayback/available?url=example.com/dead-page"

If the URL 404s in a real browser too, it is genuinely broken. If it loads fine for you, mark it as a false positive and move on.

Common mistakes

The classic errors: deleting every flagged link in bulk without checking whether the content simply moved, which strips your articles of the citations that made them credible. Treating every 403 as broken when many are just bot blocks. Fixing zero-traffic pages while your top guides stay broken. And fixing everything once, then never re-crawling, even though link rot never pauses.

FAQ