Not Found (404) in Google Search Console: What It Means and When to Fix It

Not found (404) in Search Console: what it means

If Google's indexing email brought you here, take it as good news: Google is visiting your site and reporting what it found. Not found (404) is one of the labels that email can name. The guide to Google indexing emails covers the email itself and every other reason on it.

Google's Page indexing help, re-read on October 18, 2026, says Google may have found such an address "as a link from another page, or possibly the page existed before and was deleted." Its verdict is the line to keep: "404 responses are not necessarily a problem, if the page has been removed without any replacement. If your page has moved, use a 301 redirect to the new location."

A 301 is a permanent forward: anyone who opens the old address, Google included, lands on the new page. Earlier on the same help page, in its troubleshooting section, Google narrows the to-do list to two kinds of address: it recommends fixing only missing addresses that your own site links to or lists in its sitemap.

The table sorts the common cases. To work through your own list, jump to Fine or fix, address by address; for the button in the report, see when Validate fix helps.

Why Google lists addresses that are not there

Google remembers the addresses it has met, wherever it met them, and comes back to check. The Not found (404) list is built from those return visits. Google's help says it plainly: "there is no way to tell Googlebot to permanently forget a URL, although it will crawl it less and less often." So the list shrinks on its own, slowly. The addresses come from six places:

• Your old site. Pages from before a redesign or a move to a new platform, which Google visited back then.
• An old sitemap. The file that lists your pages for Google, if it still names pages that are gone.
• Your own links. A menu item, button or footer link that points at a page you removed or renamed.
• Links on other sites. Directories, local blogs and old social posts keep linking to the addresses they knew.
• Typos. Someone meant to link to you and misspelled the address.
• Addresses nobody made. Google's May 2011 post on 404s names links mangled by copying sites or by software, and "a quick check from our side to see how your server handles unknown URLs". Its advice for those: "you can safely ignore them."

Fine or fix, address by address

This works for a site on any platform. Sign in to Search Console with the Google account the email went to, then:

1. Open Search Console and pick the site. In the left menu, click Indexing, then Pages.
2. Scroll to Why pages aren't indexed and click the Not found (404) row.
3. Read the examples. Google lists up to 1,000 addresses under the row, each with the date of its last crawl.
4. If the list is long, click Export to open it in a spreadsheet, one address to a row.
5. To see which missing addresses your own sitemap still lists, go back to the Pages report and change the dropdown above the chart from All known pages to All submitted pages. Google's help says that dropdown filters the results "by whether or not they are included in a sitemap." Any address still under Not found (404) is one your sitemap names: take it out of the sitemap, or put the page back. Then set the dropdown back to All known pages.
6. Sort each address with the table below. Most land in a row marked Fine.

A worked example: Kvalheim Bakery

Kvalheim Bakery, a bakery site built by Theo, has pages for its bake list, custom cakes, wedding cakes and holiday orders. When a site like it replaces an older one, Google often lists old addresses like the six in the drawing above a few weeks later. Here is the call on each.

Three forwards, and three addresses left alone. None of the missing addresses sits on the new pages: they come from the old site's leftovers and from links on other sites. The forward for /wedding.html lands on the page below, so a couple following the directory link reaches the wedding cakes instead of a not-found page. After the three forwards, the list still holds three correct 404s, so Validate fix stays unclicked.

Address	What it was	The call	Why
/custom-cake-orders	The old site's custom cake page	Forward (301) to /custom-cakes	It moved: the same page now lives at /custom-cakes, and Search Console shows clicks to the old address over the last 16 months.
/wedding.html	The old site's wedding cake page	Forward (301) to /wedding-cakes	Its subject lives at /wedding-cakes now. A local wedding directory also still links to it, so it shows in the Links report.
/custom-cake	Never a page: a typo	Forward (301) to /custom-cakes	URL Inspection names a neighborhood blog as the referring page. A misspelled link was meant for you.
/2023-holiday-hours	A notice from two winters ago	Leave it	No clicks, no links, and nothing replaces it.
/sample-page/	A page the old WordPress install made	Leave it	It was never in the menu and never earned a visit.
/cart.php	Never a page	Leave it	The bakery never had an online shop. Google found a stray link to it somewhere. Google's 2011 post says addresses like this can be safely ignored.

Did the old address earn its keep? Three checks in Search Console

A page that simply moved always gets a forward to its new address, whatever it earned. These three checks are for the rest. Use them on old addresses with no page on the same subject now. On a long export, they also show which addresses to handle first. Each check asks one question: did this address bring anything in? If it did, forward it to the closest page that meets the same need, not the homepage as a catch-all. If no page meets that need, leave the 404.

1. Clicks. Click Performance, then Search results. Set the date to the last 16 months, add a Page filter with the old address, and read Clicks. Google's help says Search Console keeps data for the last 16 months, so a page that only earned in one season still shows. Clicks above zero: forward it.
2. Links. Click Links, then Top linked pages under External links. Google's help for the Links report says it includes links "that Google has found over time", even to a page that "may no longer exist". An old address in that list has another site pointing at it: forward it.
3. Who sends Google there. Paste the address into the URL Inspection bar at the top of Search Console and open Page indexing. Under Discovery, Referring page is "a page that Google possibly used to discover this URL." If it is a page on your own site, change that link. If it is another site, forward the address or ask them to correct the link. An empty Referring page, Google adds, "doesn't mean that no referring page exists".

The four fixes, and which one fits

Every address on the list ends in one of four places. The drawing runs the questions in order, and the first yes decides. An address with no yes keeps its 404, which is the right answer for it.

1. Restore it. The page should be live and went missing by mistake: unpublished, deleted or renamed by accident. Put it back at the same address.
2. Forward it (301) to the page on the same subject. One old address to one new page. Google's Page indexing help gives the rule in one line: "If your page has moved, use a 301 redirect to the new location." Platforms call this a redirect, and the redirect table in the redesign guide shows where the setting lives on each major website builder. On Theo, SEO repair on Starter and Pro forwards old addresses that still get search traffic to their matching new pages, and on any plan you can ask in chat for a forward.
3. Fix your own link. If your menu, a button or your sitemap points at the address, change the link to the live page, or take the address out of the sitemap.
4. Leave the 404. For a page removed on purpose with nothing to replace it, or an address nobody made. The 2011 post says Google would "actually prefer that, when you get rid of a page on your site, you make sure that it returns a proper 404 or 410 response code".

Validate fix: only when every address should show a page

Validate fix asks Google to recheck every address listed under Not found (404). On the whole list, it helps in one case only: every listed address should show a page, and now does, because you restored it or forwarded it.

If some of the addresses are meant to stay gone, skip it. Google's help describes each recheck: "If the instance is still present, issue state changes to Failed and validation ends." A correct 404 is still a 404 when Google comes back, so the first one it rechecks ends the validation. Skipping it costs nothing, because "Google updates your instance count whenever it crawls a page with known issues, whether or not you explicitly requested fix validation."

If you put back only a few pages, Google's help has a way to check just those, in a tip called Validate your fixes by sitemap. Submit a sitemap that lists only those pages, pick it in the dropdown above the chart, then click Validate fix. In Google's words, "the validation will apply only to items in the sitemap at the time you requested validation." Making that extra sitemap is a job for whoever runs your website.

When it does fit, click it once and wait. Google's help: "Validation typically takes up to about two weeks, but in some cases can take much longer, so please be patient." Google emails you when it passes or fails. Every state of the timeline is drawn in the Search Console email guide.

Do 404s hurt SEO?

No. A 404 on an address that has nothing to show does not count against the rest of the site, and Google has said so in nearly the same words for fifteen years. Here is the record, each line re-read on October 18, 2026.

Not found (404), Soft 404 and Page with redirect: how they differ

Google's email often names more than one reason. This table tells the close ones apart, with the guide for each.

theo.site's 8 retired addresses

theo.site was rebuilt on September 30, 2026. Its redirect map forwarded 49 old addresses to the closest new page and retired 8 with nothing to replace them: a blog section of four job descriptions and its index page, a post on social media jobs, and a food truck business plan and its index page. None of them was about websites, and no new page covered the same subject.

None of the 8 was forwarded, because a many-to-one forward to the blog would be a soft 404. Each answers 404, which Google treats the same as 410. On October 18, 2026, each address answered 404 in a live check, and Search Console's URL Inspection showed what Google held for each one.

Six of the 8 were unknown to Google: it had never visited them, so they cannot appear in the report until it does. One, the social media jobs post, sits under Crawled - currently not indexed from a June 14, 2026 visit (Pacific time), before the rebuild. One, /blog/business-plans, was filed as Not found (404) after a visit on October 17. The answer for every row is to leave it: no forward, no Validate fix, no Removals request. A well-kept site still has a Not found (404) row. What matters is that each address on it was retired on purpose.

What a visitor sees at a retired address

A good not-found page helps people without hiding anything from Google. theo.site's shows a short note, the menu and links to the main pages, while the address still answers with a 404 status, so Google knows the page is gone. An address that does not exist on the bakery site from the example above answers the same way, shown next.

The same check on a site built by Theo

Theo builds each site with its own not-found page. Kvalheim Bakery's says "That page isn't in the oven." and offers a way back to the bakery, while the address still answered with a 404 status. An address like it lands under Not found (404) if Google ever asks for it, not under Soft 404.

What to send whoever runs your website

Owners on Theo can forward Google's email to Theo support, or ask in chat, instead.

Frequently asked questions