A soft 404 is a web page that tells a visitor the thing they asked for is missing, or shows almost nothing at all, while the server reports the request as a success. The page says "not found"; the HTTP status line says 200 OK. Search engines read status codes first, so a soft 404 leaves them to work out from the content that the page is effectively an error. Google, Bing and other crawlers run classifiers to do exactly that. The term exists because the mismatch is common, costly to crawlers and a source of quiet failures for ads and product listings.

Status codes and the gap a soft 404 fills

Every response a web server sends begins with a three-digit status code. The current definitions sit in RFC 9110, the Internet Engineering Task Force (IETF) document on HTTP semantics published in June 2022. 404 Not Found means the server "did not find a current representation for the target resource or is not willing to disclose that one exists". 410 Gone goes further, signalling that the resource is no longer available and that "this condition is likely to be permanent". 200 OK means the request succeeded.

A soft 404 is a page that should carry a 404 or 410 but carries a 200. Google's 2010 announcement put it plainly: a soft 404 "occurs when a webserver responds with a 200 OK HTTP response code for a page that doesn't exist". Google's current help documentation widens it to include redirecting users to another page, such as the homepage, instead of returning a 404.

The distinction from a hard 404 is about who decides: the server, or the search engine after reading the page.

How soft 404s are produced

Most soft 404s are accidents of configuration rather than design.

  • Custom error templates served with 200. A content management system renders a friendly "page not found" template without setting the status code.
  • Redirects of removed URLs. Deleted pages are sent by 301 or 302 to the homepage or a broad category page rather than an equivalent.
  • Empty or near-empty pages. Category pages with zero products, internal search results with no matches and expired classified listings.
  • Out-of-stock product pages. Product pages stripped down to "currently unavailable" with no description or specifications.
  • Single-page applications. Client-side JavaScript frameworks typically use a catch-all route that serves the same HTML shell, with a 200, for every URL, and the browser then decides whether to show content or a "not found" view. Google's 2018 codelab on "sneaky 404s", written by Martin Splitt, taught developers to add exactly that wildcard route so deep links would load.
  • Challenge pages. CAPTCHA interstitials that return 200 to a crawler.

How search engines detect them

Because the status code cannot be trusted, detection is a content classification problem. Bing described its approach in March 2012, when Richard Qian of Bing R&D wrote that its classifiers "use page content such as key phrases in the page's title, body and URL to determine if the page is a soft 404".

Google has not published its signals. At Search Central Live Asia Pacific in July 2025, Gary Illyes described the process in general terms, according to reporting by Search Engine Journal based on a LinkedIn summary by Kenichi Suzuki: when Google detects patterns that signal a missing or low-value page, it flags the URL as a soft 404. For JavaScript sites the check happens after rendering. Google's documentation, clarified in December 2025, says pages returning a non-200 status may not be rendered at all, so a 200 is what sends a single-page app's error view into the rendering queue to be judged.

Once flagged, the URL is excluded from the index in the same way as a true 404. In Google Search Console, it appears in the Page indexing report under the reason "Soft 404", and the URL Inspection tool shows the same verdict for an individual address.

Fixes and their trade-offs

Google's 2010 post framed three outcomes for each flagged URL, and the guidance has not changed since.

  1. Return a real 404 or 410 when content is gone and has no replacement. Google's help page says to "let the old URL return a 404 or 410" and not to "create fake content, redirect to your homepage, or use robots.txt to block 404s".
  2. Redirect with a 301 when the content has moved to a specific equivalent page.
  3. Keep the 200 and make the page substantive when it is a false positive, for example by adding product detail to an out-of-stock page.

For single-page applications, where Google's JavaScript documentation calls meaningful status codes "impossible or impractical", two workarounds are documented: a JavaScript redirect to a server URL that returns a genuine 404, or a JavaScript-injected noindex robots meta tag on the error view. PPC Land's analysis of the December 2025 changes judged the redirect safer, because Google may skip rendering on pages that carry noindex.

Crawl budget and indexing effects

The direct cost of a soft 404 falls on crawling. Maile Ohye's August 12, 2008 post "Farewell to soft 404s" warned that search engines "may spend much of their time crawling and indexing non-existent, often duplicative URLs", so that genuine pages "may not be discovered as quickly or visited as frequently". Google's crawl budget documentation, last updated July 22, 2026, is more blunt: "soft 404 pages will continue to be crawled, and waste your budget", while a 404 "is a strong signal not to crawl that URL again".

That asymmetry matters only for some sites. The same document says crawl budget is a concern mainly for sites with more than 1 million unique pages changing weekly, or more than 10,000 changing daily. For a small site, a handful of soft 404s is a reporting nuisance rather than a constraint, as PPC Land's explainer on crawl budget sets out.

Rankings are a separate question. In August 2024, John Mueller said 404 errors do not affect the ranking of other pages and warned against redirecting deleted pages to the homepage.

Why it matters for advertisers and merchants

Paid channels apply their own version of the same test. Google Ads' "Destination not working" policy lists HTTP errors such as 404 and 500, but also pages that return content like "Whoops! Nothing to see here!" or "Site Under Construction". That is a soft 404 standard applied to landing pages: a 200 does not protect an ad if the page reads as an error. The ad-review crawler behind those checks is Google AdsBot.

Merchant Center routinely crawls product landing pages and disables products whose pages are unavailable, stopping them "showing in ads and cost-free listings" until access is restored. PPC Land has covered how that flag stops Shopping ads even when Google's crawler is at fault.

Out-of-stock pages sit where the two systems pull in different directions. Merchant Center requires the page to stay live and show the item as unavailable; in March 2026 Google tightened the rule so the buy button must be greyed out rather than hidden. Search, meanwhile, may treat a thin out-of-stock page as a soft 404. In April 2022 Mueller said Google may "assume it's more like a soft 404 error" and "drop that URL from the search results", while adding that a page with substantial product information "can still be quite relevant". The page that satisfies both is a full product page with an accurate availability label.

Limitations and disputes

The central weakness is that detection is a guess. False positives are common on legitimately short pages: store locator entries, glossary definitions, image pages, and product pages whose text says "sold out" in a prominent heading. Owners cannot overrule the classifier except by changing the page.

Reporting has also shifted. On January 11, 2021, Google announced that Search Console's soft 404 reporting was "now more accurate". Glenn Gabe of G-Squared Interactive found that many URLs previously flagged moved into other statuses, including "Crawled - currently not indexed", and some former soft 404s were indexed. Trend lines across that date are not comparable.

The crawl budget argument is contested in scale rather than direction, since Google's own thresholds exclude most sites. Engines also differ. Bing documents a content classifier; other crawlers, including many AI crawlers and SEO tools, still take a 200 at face value, so a soft 404 that Google drops can persist elsewhere.

  • 404 Not Found. A server-declared status. A soft 404 is an engine's inference about a page with a 200.
  • 410 Gone. Explicitly permanent removal. Google says it treats all 4xx codes except 429 the same way for indexing, so the practical difference from 404 is small.
  • 301 and 302 redirects. Legitimate moves to an equivalent page. A redirect to an unrelated page can itself be classified as a soft 404.
  • Thin content and noindex. Thin content is a quality judgement about pages meant to exist; noindex is an owner's instruction. A soft 404 is the engine concluding the page should not exist at all.

Recent developments

Soft 404s have moved up Google's list of crawling problems. At Search Central Live Deep Dive Europe in Barcelona, which ended October 2, 2026, Illyes described soft 404s as one of two growing problems, citing challenge pages that return 200 OK with no real content, according to an agency recap of the event. In July 2026, a classifieds site owner reported that over half of Googlebot's hits went to 404 pages, and the same month Google dropped its 2007 rule on blocking internal search pages, though Mueller still described indexed search pages as "very inefficient".

Timeline

  • May 1996 - RFC 1945 (HTTP/1.0) defines 404 Not Found.
  • January 1997 - RFC 2068 (HTTP/1.1) introduces 410 Gone.
  • May 2004 - Ziv Bar-Yossef, Andrei Broder, Ravi Kumar and Andrew Tomkins describe "soft 404s" in the paper "Sic transit gloria telae" at the 13th International World Wide Web Conference.
  • August 12, 2008 - Google's Maile Ohye publishes "Farewell to soft 404s".
  • June 7, 2010 - Google Webmaster Tools begins reporting soft 404s under Crawl errors.
  • March 19, 2012 - Bing describes its soft 404 classifiers.
  • January 2018 - Google rolls out the redesigned Search Console with the Index Coverage report.
  • January 11, 2021 - Google says Search Console soft 404 reporting is "now more accurate".
  • April 2022 - Mueller says thin out-of-stock pages may be treated as soft 404s.
  • June 2022 - RFC 9110 replaces earlier HTTP semantics specifications.
  • August 2024 - Mueller says 404s do not hurt rankings and warns against homepage redirects.
  • July 2025 - Illyes discusses soft 404 detection and crawl cost at Search Central Live Asia Pacific.
  • December 18, 2025 - Google clarifies that non-200 pages may not be rendered.
  • March 2026 - Merchant Center tightens out-of-stock buy button requirements.
  • July 22, 2026 - Google updates its crawl budget documentation, which says soft 404s waste budget.
  • October 2, 2026 - Illyes calls soft 404s a growing problem in Barcelona.

Summary

Who. Website owners and developers create soft 404s, usually by accident. Google, Bing and other search engines detect them with content classifiers, and Google reports them in Search Console. Advertisers and merchants meet the same logic through Google Ads and Merchant Center landing page checks.

What. A soft 404 is a page that returns HTTP 200 OK while its content says the page does not exist or offers no meaningful content. The search engine overrides the status code and treats the URL as an error.

When. Researchers described the phenomenon in May 2004. Google named it publicly in August 2008, began reporting it in June 2010 and has revised its reporting and JavaScript guidance since, most recently in December 2025 and July 2026.

Where. Soft 404s occur on error templates, redirected removed URLs, empty category and search pages, thin out-of-stock pages and JavaScript single-page applications, and surface in Search Console's Page indexing report and URL Inspection tool.

Why. Status codes are how crawlers decide what to fetch and index. When a page lies about its status, search engines waste crawl capacity rechecking it and must guess, sometimes wrongly, which pages deserve to be indexed.