Updated
A soft 404 is a page that tells visitors it doesn’t exist, or has almost nothing on it, while the server returns 200 (OK) instead of a 404 status. Search Console lists these under Soft 404 in the Page indexing report, and Google leaves them out of search. Fix each one according to what it really is: return 404 or 410 if the content is gone, redirect if it moved, or make the page load real content if it should exist.
What is a soft 404 error?
Google’s crawling errors guide defines a soft 404 as a URL that returns a page telling the user the page doesn’t exist, together with a 200 (success) status code, and notes it can also be a page with no main content. Google decides this from the content, not the status code. Its examples:
- A missing server-side include file.
- A broken connection to the database.
- An empty internal search result page.
- A JavaScript file that didn’t load.
Google calls this a bad user experience: visitors think the page works, then see an error. Such pages are excluded from Search.
| Soft 404 | Not found (404) | Gone (410) | |
|---|---|---|---|
| Status code | 200, or another success code | 404 | 410 |
| What the page says | “Not found”, “no results”, or nearly nothing | Not found | Gone |
| Indexed? | No | No; removed if it was indexed | No; removed if it was indexed |
| Search Console reason | Soft 404 | Not found (404) | Listed under a 4xx reason |
| Is it a problem? | Yes: the status is wrong or the page broke | Not if the page is gone on purpose | Not if the page is gone on purpose |
Google treats all 4xx codes except 429 the same way: it tells the indexing system the content doesn’t exist, removes the URL if it was indexed, and crawls it less and less (HTTP status codes and Google). A 410 is still the more accurate code for a page removed for good.
How to fix a soft 404
Inspect an example URL in Search Console, click Test live URL, then View tested page to see the screenshot and the HTTP response Google got. Then pick the case that fits, following Google’s three cases:
The page and content are gone
Return 404 (not found) or 410 (gone). Keep a helpful custom 404 page for visitors (search box, popular categories, a link home), but make sure the server sends the 404 status with it. You can check the code yourself:
curl -s -o /dev/null -w "%{http_code}\n" https://example.com/listing/closed-business
# 200 here, on a page that says "listing not found", is a soft 404The page moved, or has a clear replacement
Return a 301 redirect to the new URL; it will then be reported as Page with redirect, which is fine. Only redirect to a page that really replaces the old one. Sending hundreds of removed pages to the home page or one category doesn’t give visitors what they came for; a 404 or 410 is the honest answer when nothing replaces the page.
The page should exist and has content
If a real page is flagged, Google likely didn’t see its content: it didn’t load properly for Googlebot, critical resources were missing or blocked by robots.txt, or it showed a prominent error while rendering. Look at the screenshot in the live test. If it is blank or shows an error, fix whatever the page depends on: an API that times out, a script blocked in robots.txt, a database error, a page that only fills in after a user action.
On JavaScript sites that render pages in the browser, Google’s JavaScript SEO basics suggests two ways to avoid soft 404s: redirect error views to a URL that returns a real 404, or add a <meta name="robots" content="noindex"> tag to error views with JavaScript.
After the fix, open Soft 404 in the Page indexing report and click Validate fix. The page indexing issues guide explains how validation runs.
Soft 404s on directory websites
Directories create soft 404s in bulk because their pages are generated from data, and data runs out. These are the usual sources and what to do with each:
| Page | Why Google sees an error | What to return |
|---|---|---|
Empty category-in-city page: /plumbers/hutto-tx with “0 plumbers found” | The main content is a “no results” message. | Don’t publish the page until it has listings (404 until then), or noindex it below a minimum count. If you have real nearby listings, label them honestly as nearby and add local detail; a page of useful content is not a soft 404. |
| Removed or expired listing: “This listing is no longer available” | The page says the content is gone, with a 200. | 410 if it is gone for good, 404 otherwise. 301 only if the business moved to another listing on your site. |
| Unpaid or lapsed paid listing hidden behind a placeholder | A near-empty page. | Downgrade to a free listing that still shows the business’s details, or 404 it. Don’t leave a placeholder at 200. |
| Closed business kept as a record | Little content and a “permanently closed” notice. | Keep it at 200 only if the page still helps searchers (address history, reviews, similar businesses nearby). Otherwise 410. |
| Expired deals, events and job posts | An “ended” page with nothing else. | 404 or 410 once ended, or keep a real page with the details and related live items. |
Pages past the end of a list: /plumbers/austin-tx?page=40 | An empty list. | 404 for page numbers beyond the last page. |
Internal search with no results: /search?q=xyz | An empty result page. | Noindex search results or block them in robots.txt. They shouldn’t be indexable at all. |
| Listing pages that load from an API in the browser | The data call failed or was slow for Googlebot, so the page rendered empty. | Render listing content in the HTML, or fix the API; check the live test screenshot. |
Fix these in the template, not page by page: a rule like “return 404 when a category-in-city query has zero listings” clears thousands at once. Google’s guide also notes that soft 404 pages waste crawling that could go to useful pages, which matters on a big directory (see discovered, currently not indexed).
How to find every soft 404 on a large site
The report’s example list stops at 1,000 URLs. To see the full extent:
- Filter the Page indexing report by sitemap, one sitemap per page type, and open Soft 404 in each. The counts tell you which template produces them.
- Export the examples and check what your server actually returns for each. A simple loop prints the status code next to every URL:
while read -r u; do
printf "%s %s\n" "$(curl -s -o /dev/null -w '%{http_code}' "$u")" "$u"
done < soft-404-urls.txtEvery line that starts with 200 is a page the server still calls fine. Group them by URL pattern, then open one of each pattern in URL Inspection’s live test to see the screenshot Google got.
Check your server logs for Googlebot requests that got a 200 and a tiny response size. Empty category pages and placeholder listings are usually far smaller than real ones.
Prevent soft 404s in your templates
- Zero-result guard. A category, city or category-in-city template that finds no listings returns 404, or renders with noindex, instead of a “no results” page at 200.
- A removal workflow. Deleting or expiring a listing sets 410, or a 301 when you pick a replacement. It never leaves a placeholder.
- Bounded pagination. Page numbers past the last page return 404.
- Real errors get real codes. When the database or an API fails, return a 5xx error such as 503, not an apology page with a 200. Google treats a short spell of 503s as temporary and comes back; a 200 error page invites a soft 404.
- Content in the HTML. Render listing details on the server, so a slow API call can’t leave Googlebot with an empty page.
Is error 404 permanent?
No. A 404 is just the answer the server gave on that request. If the page comes back with a 200, Google can index it again. Google keeps trying URLs that return 404 for a while, in case the error is temporary, and crawls them less and less often; there is no way to make Google permanently forget a URL. To keep the list readable, the Page indexing report only shows URLs that returned 404 in the past month (Page indexing help).
A 404 for a page you removed on purpose isn’t a problem. Google’s help recommends fixing only the 404s you link to yourself or list in a sitemap, and redirecting pages that moved. More in fixing 404 errors in Search Console.
What is 404 Not Found?
404 Not Found is the HTTP status code a server sends when it has nothing at the requested URL. The browser shows the site’s error page; Google takes the 404 as a sign the content doesn’t exist and doesn’t index it. A soft 404 is the same situation with the wrong code: the visitor sees “not found”, the server says “200 OK”.
Find soft 404s across a directory with DirectoryLens
Search Console shows up to 1,000 example soft 404 URLs. DirectoryLens’s Google indexing report inspects your pages daily by page type, and its Individual pages tab shows Google’s status for each URL, so you can see whether soft 404s cluster in one template, such as category-in-city pages.
With the optional light crawl on, Thin and unused pages reads the status code and listing count of sampled pages. It recommends a 410 or a redirect for pages that return errors, and noindex or more listings for category and city pages below your thin-page threshold with no impressions. It also lists listing pages with no impressions in 90 days, the ones most likely to be expired or empty, with a suggested action. DirectoryLens never changes your site; it tells you which pages to look at.