How a 404 page works
Two things happen when someone requests an address your site does not have. The server answers with the status code 404, meaning not found, and it sends some page for the browser to display. The code is the part machines read; the page is the part people read. A custom 404 page replaces the bare server message with something that carries your design, your navigation and a way forward.
The pairing matters more than the design. A page that looks like an apology but answers 200 is a soft 404: search engines are told the address exists and holds real content, so they may index it and keep returning to it. Where a page has been deliberately removed for good and will never come back, 410 is the more precise answer, but a plain 404 is perfectly acceptable.
Why the 404 page matters
Dead addresses are unavoidable. Products are discontinued, campaigns end, someone mistypes a link in a newsletter, another site links to you with a stray character on the end. The question is not whether people will land on missing pages but what happens when they do.
A blank server error reads as a broken site, and most visitors leave rather than hunt through a menu. A 404 page that explains the situation plainly and offers the routes people actually want — the main sections, a search box, a contact link — recovers a share of those visits. It is also a diagnostic surface: the addresses that keep appearing in your not-found reports tell you which links are broken and which URLs the outside world still expects to work.
Common mistakes with 404 pages
The most common is redirecting every missing address to the home page. It hides the problem instead of solving it, drops visitors somewhere unrelated, and leaves you with no record of what was broken. Redirect when there is a genuine replacement page; otherwise let the 404 do its job.
Next is the decorative 404: a large graphic, a joke, and no way to reach anything. Humour is fine, but the page still has to answer the visitor’s question, which is where do I go now. Then there is the soft 404 described earlier, which happens easily when a page builder or a plugin generates the not-found page as ordinary content. And many sites never check their not-found reports at all, so genuinely broken internal links survive for years.
How to act on it
Confirm the status code first. Request a deliberately wrong address on your own site and check the response is 404 rather than 200 — your browser’s network tab or Search Console will tell you. Then make the page useful: say clearly that the address does not exist, keep the full header and footer so navigation still works, offer search, and point to the few destinations most visitors want.
Review the not-found reports in Search Console and your analytics regularly. Internal links pointing at missing pages should be fixed at the source. External links pointing at addresses that once existed deserve a redirect to the closest real page, since that link value is otherwise wasted. Everything else can safely stay a 404, and keeping that distinction clean is part of routine technical SEO maintenance.
