Summary: A 404 is a page that no longer exists, and a handful of them is normal on any site with history. They become a problem when they sit on pages people actually reach, when they are listed in your own sitemap, or when a rebuild moved every URL without redirects. This guide covers what a 404 actually costs you, how to find yours in Search Console and by crawling your own sitemap, when to redirect versus when to let a page die, and why a useful 404 page recovers visitors a default one loses.
What a 404 actually costs
Two things, and they are not equally serious.
The visitor. Someone clicked a link expecting something specific and got an error. On most default 404 pages there is nothing to do next, so they leave. If that link was in a Google result or an email you sent, you paid for that click in one way or another.
Crawl budget. Google allocates a finite amount of attention to your site. Every dead URL it fetches is attention not spent on a page that matters. On a small site this is irrelevant; on a site with hundreds of pages where most are not indexed, it is not.
What a 404 does not do is directly penalise you. Google expects pages to disappear. A few 404s are normal and healthy.
Where to find yours
Search Console is the first place: Pages, then “Not found (404)”. This lists what Google has actually tried to fetch, which is the set that matters. It also tells you where the link was found.
Your own sitemap is the second, and the more embarrassing one. A dead URL in your sitemap is your site actively instructing Google to go and fetch something that is not there. Fetch your sitemap.xml, pull out every URL, and check each one returns 200. This is a ten-minute job and it finds problems Search Console has not got to yet.
Your own pages are the third. Internal links to pages you deleted years ago are common, and they are entirely within your control to fix.
Redirect, restore, or let it die
Redirect when there is a genuine replacement. A 301 to the closest equivalent page passes on most of whatever standing the old URL had, and sends the visitor somewhere useful. This is right when a page moved, when two pages merged, or when a URL changed in a rebuild.
Restore when the page was deleted by mistake and people are still looking for it.
Let it die when there is no replacement. Redirecting an unrelated page to your homepage is worse than a 404: Google treats it as a soft 404 anyway, and the visitor gets dumped somewhere they did not ask for. If a page is genuinely gone, a clean 404 is the honest answer.
The mistake worth avoiding is the blanket redirect — sending every missing page to the homepage. It looks tidy and it helps nobody.
Rebuilds are where 404s come from in bulk
One or two broken links accumulate slowly. Hundreds appear overnight, and it is nearly always the same cause: a site was rebuilt and the URLs changed.
Every page that changes address needs a 301 from the old one. If that mapping is not part of the rebuild plan, the new site launches and every ranking, every bookmark and every inbound link points at nothing. We are called in to reconstruct these afterwards, which is harder than doing it at the time because the old URLs have to be recovered from Search Console and archives.
If you are planning a redesign, ask what happens to the URLs before anything else.
Make your 404 page do something
Most 404 pages are a dead end. Yours does not have to be. A useful one says plainly that the page has gone, offers a search box, links to your main sections, and gives a way to contact you. That turns some proportion of a lost visit into a recovered one, and it costs an hour to build.
Two things to avoid: a 404 that returns a 200 status code, which tells Google the error page is real content and gets it indexed; and a cute 404 with no navigation, which is a dead end wearing a joke.
How often to check
Quarterly is enough for most business sites, and after any rebuild, migration or bulk deletion. Regular website maintenance should include it, along with the internal links that break quietly when you rename a page.
Common questions
Do 404 errors hurt my Google rankings?
Not directly. Google expects pages to disappear and a few 404s are normal. They become a problem indirectly — when they waste crawl budget on a large site, when they sit on URLs that still have inbound links, or when a rebuild broke every address at once.
Should I redirect every 404 to my homepage?
No. Redirect only where there is a genuine equivalent page. A blanket redirect to the homepage is treated as a soft 404 by Google and is worse for the visitor than an honest error page, because they get dumped somewhere they did not ask for.
How do I find the 404s on my website?
Start with Search Console under Pages, then "Not found (404)" — that is what Google has actually tried to fetch. Then check your own sitemap by pulling out every URL and confirming each returns 200. A dead URL in your own sitemap is the worst kind because you are asking Google to crawl it.
What should a good 404 page include?
A plain statement that the page is gone, a search box, links to your main sections and a way to contact you. It should also return an actual 404 status code — an error page that returns 200 gets indexed as if it were real content.
Want to know what is broken on your site?
We will crawl it, list every dead link and every 404 in your sitemap, and tell you which ones are worth redirecting.
Get in touch Call (647) 385-5532
BridgeWeb has been building websites in North York, Toronto since 2006 — 115 projects across 14 industries.
