A 404 error means a requested URL doesn't exist on the server, the honest, correct response when content is genuinely gone. That makes 404s different from most technical SEO problems: the status code itself isn't the bug. The problem is always one of two specific situations, and they need opposite fixes.
The Two Situations, and Why They're Different
Situation one: something still links to a page that's gone. A product was discontinued, a blog post was renamed, a page moved during a redesign, and a link somewhere, on your own site or someone else's, still points to the old URL. This is fixable and worth fixing: a 301 redirect to the closest relevant replacement recovers the traffic and the link equity that URL had built up.
Situation two: a visitor or bot requests a URL that never legitimately existed. A typo in the address bar, an old bookmark to a page from a site redesign three owners ago, a scanner probing for /wp-admin on a site that isn't even WordPress. This is not a problem. A clean 404 is the correct response here, and trying to redirect every conceivable bad URL to somewhere plausible creates more mess than it solves.
When It's Actually Worth Fixing
- Internal links on your own site point to the dead URL. No excuse for this one, it's fully within your control. See our technical SEO checklist for the wider crawlability review this sits inside.
- The URL has real backlinks pointing to it, visible in your analytics as steady referral traffic that suddenly stopped. That's inbound authority worth redirecting to a live page rather than letting it dead-end.
- It's a page from a recent site migration or redesign, where the old URL structure is still fresh in search results and bookmarks. See our website migration checklist for handling this at scale rather than one URL at a time.
When It's Genuinely Fine to Leave a 404
- Random bot probes and scanner traffic aimed at URLs your site never had, this is normal background noise on any live website and not worth a redirect rule.
- A URL nobody links to and nothing in Search Console flags, a 404 that no crawler or visitor is actually hitting isn't costing you anything.
The Redirect-Chain Trap
Fixing a 404 badly is worse than leaving it. Redirecting /old-page to /newer-page, which itself later gets redirected to /current-page, creates a redirect chain, each hop adds latency and, past a couple of links, some crawlers stop following the chain entirely. Always point a fresh redirect straight at the final, current URL, not at another redirect. See our 301 vs 302 redirects guide for getting the redirect type right in the first place.
How to Check What You Actually Have
- Google Search Console's Page Indexing report lists "Not found (404)" URLs Google has encountered, cross-reference against your own internal links before assuming any given one needs fixing.
- Check your analytics for referral traffic to 404s specifically, that's the signal a URL is worth a redirect rather than a shrug.
- Crawl your own site for internal links to non-200 URLs rather than waiting for Search Console to catch them, this is the fully-within-your-control subset from the list above.
The Fast Way to Check Yours
Distinguishing a 404 worth fixing from one that's fine to leave, across an entire site, is exactly the kind of thing that doesn't scale by hand. AuditCrow's free scan flags broken links and status-code issues automatically. If your traffic drop looks bigger than a handful of dead links would explain, see our guide on why your website isn't showing up on Google for the fuller diagnostic list.