GuidesIndexed

Soft 404 errors: a missing page that says it’s fine

Your server tells Google a page loaded normally, and the page itself looks empty or gone. Search Console calls that a soft 404. On service business sites it usually comes from a jobs or search page, a redirect setting or an empty archive.

John Geater, founder of IndexFirst6 min read
The short answer

A soft 404 is an address that returns 200 OK even though the page looks missing, blank or like an error. Google keeps it out of search. The fix is to make the status code tell the truth: if the page should not exist, return a proper 404 or 410; if it should, give it real content.

What counts as a soft 404?

Two things have to happen together. The server answers 200 OK, which means “this is the page you asked for.” And what loads does not look like a page. Google’s status code documentation says Search Console shows a soft 404 when the content suggests “an empty page or an error message.” A friendly “sorry, we can’t find that” message sent without a 404 code counts. So does forwarding dead links to the homepage: Google’s help page on 404s warns you not to “redirect to your homepage” for missing pages.

What changes is how Google treats the URL afterwards. A genuine 404 is dropped and, in Google’s words, “the crawling frequency gradually decreases.” A soft 404 is left out of the index as well, but Google keeps fetching it. Its crawl budget guide says these pages “will continue to be crawled, and waste your budget.”

A proper 404: page and status agreeExample: example.co.uk/services/gdpr-audit-2019 was removed. The browser shows “We can’t find that page” with a link to the services list. The server sends 404 Not Found. Google drops the URL and visits it less often.Proper 404example.co.uk/gdpr-audit-2019Page not hereWe can’t find that page.See our servicesStatus sent404 Not FoundGoogleDrops it, visits less oftenA soft 404: the homepage, sent with 200 OKExample: example.co.uk/services/gdpr-audit-2019 was removed. The browser shows the Example Consulting homepage instead. The server sends 200 OK. Google keeps the URL out of its index and carries on fetching it.Soft 404example.co.uk/gdpr-audit-2019Example ConsultingStatus sent200 OKGoogleLeaves it out, keeps fetching
One removed page on Example Consulting, a made-up consultancy, handled two ways. Based on Google’s documentation on HTTP status codes and crawl budget.

Where do service business sites pick up soft 404s?

A handful of places come up again and again. A recruitment agency’s jobs page that shows “no vacancies found” for any search Google tries. A freight firm’s tracking page that does the same for any consignment reference. An office page left with only an address after the office closed. WordPress tag and author archives with one post, or none. A quote form that returns an empty results page when the parameters make no sense. And a redirect plugin set to send every missing URL to the homepage.

Each one is a page that answers 200 OK while saying, in effect, that nothing is here. The redirect setting is the one to find first, because it turns every dead link anyone has ever pointed at your site into a soft 404 at a stroke.

How do you fix a soft 404?

Work out where the 200 is coming from. On most service business sites it is one of these five.

A plugin or theme sends dead links home

Some WordPress redirect plugins, and some themes, have an option to forward every 404 to the homepage. It looks tidy and it is exactly what Google’s help page asks you not to do. Turn the option off and let missing addresses show your 404 page with a 404 status. Where a removed page has a genuine successor, say an old GDPR audit page replaced by a new data protection page, add a 301 to that page instead. If nothing replaces it, a 404 or a 410 is fine: Google “treats 410s (Gone) the same as 404s.”

Then check the result from a terminal. The first line back should show 404:

In a terminal

curl -I https://www.example.co.uk/no-such-page

HTTP/2 404

curl -I asks for the headers only. Any made-up address on your own domain will do. If the first line says 200, or a 301 that ends on the homepage, the setting is still on.

The jobs page answers every search

A URL like /jobs/12345 or /jobs?ref=12345 that loads a “this vacancy is no longer available” message with 200 OK is a soft 404 for any job that has closed or never existed. A consignment tracking page on a freight site does the same for unknown references. If your jobs board or tracking system can, have it return a 404 or 410 for those. If it cannot, add noindex to the results template and keep those links out of your sitemap and menus. The search or lookup form itself can stay indexable.

The page is decided in the browser

Some job-search widgets, client portals and booking tools load an empty shell and fill it in with JavaScript, so the server has already sent 200 before anyone knows the page is empty. Google’s JavaScript guide offers two ways out: “use a JavaScript redirect to a URL for which the server responds with a 404 HTTP status code,” or add <meta name="robots" content="noindex"> to the error state with JavaScript. The redirect is the cleaner of the two, because Google gets a real 404.

A removed service page forwards to the homepage

You stopped offering asbestos surveys, deleted the page and pointed it at the homepage. Search Console can report that as a soft 404 as well, because the homepage is not a replacement for what was there. Point it at the closest real page if there is one, perhaps your general building surveys page. If there is none, let it return 404 or 410.

The page has nothing on it

An empty vacancies page, a tag archive with no posts, an office page with a postcode and nothing else. Google reads these as empty. Either give the page something useful (opening hours, the services run from that office, a contact) or add noindex until it has content. For a closed office with no successor, remove the page and let it 404.

After the fix, check one of the flagged URLs with the live test in URL Inspection and look at the screenshot of what Google rendered. Then press Validate fix on the Soft 404 row in the report. Google says validation “typically takes up to about two weeks.” If a page then moves to a different status, our guides to crawled - currently not indexed and discovered - currently not indexed carry on from there.

Which soft 404s can you ignore?

Most of the ones for pages that really are gone. That is the report working as intended. Google says “404s don’t harm your site’s indexing or ranking,” and the crawl budget warning comes from a guide written for very large sites. A few dead URLs on a 60-page accountancy site cost next to nothing.

Two kinds deserve your time. The first is a page you want customers to find, such as a location page or a service page. A soft 404 keeps it out of Google Search and out of AI Overviews, because Google only uses pages that are “indexed and eligible to be shown in Google Search with a snippet.” The second is a site-wide setting, because one switch turns every mistyped link into a page Google keeps coming back to. If a full, genuine page appears on the list, run the live test before rewriting anything: check that Google actually saw the content you see.

Google and AI tools have to know which of your pages exist before they can recommend any of them. Clearing up status codes and soft 404s is one of the first jobs we do on every client’s site, on one plan at £1,750 a month. See what the plan includes.