What does it mean to be deindexed?

Being deindexed means a page Google had in its index is no longer in it. De-indexing is the removal itself: Google discards its stored copy of the URL, and the page stops existing as far as search is concerned. It still loads for you and still sits in your sitemap, but it returns no site: result, earns no impressions, and ranks for nothing.

Two other problems get mistaken for it. A page that never got indexed was never in. A page that ranks on position 40 is in, and losing. Deindexed is the third case, and the only one where you know for certain the page cleared Google’s bar once. That single fact changes the diagnosis: something changed, and it was probably you.

So “Google deindexed my site” is worth testing before you accept it. Most of the time the pages are still indexed and something else went wrong, and the checks below separate the two in about a minute. This is SEO strategy work: what you do next depends entirely on which of the four states your pages are actually in.

If you landed here wanting to remove your own content from Google, that is the opposite job, and a legitimate one. Skip to how to deindex a page on purpose.

If pages dropped out and you need the cause found fast, that is my Google penalty recovery service.

The four Google indexing states, and what each one means

Google Search Console reports four different states that most people compress into “not indexed,” and each one names a different failure with a different fix.

URL is unknown to Google means there is no record of the URL. Nothing to debug on the page itself, because Google has never been there. Either it has not read your sitemap since you published, or it has dropped the URL from its queue.

Discovered - currently not indexed means Google knows the URL exists, usually from your sitemap, and has not fetched it yet. This is a scheduling decision. Google’s help text says it may postpone the crawl to avoid overloading your server. On a site that answers in a few hundred milliseconds, the likelier reading is that Google weighed your URL against everything else it could crawl and picked something else.

Crawled - currently not indexed means Google fetched the page, read it, and declined to index it. That one is a quality verdict, and it is its own problem with its own fix.

Indexed means it is in.

The states form a ladder, and a page climbs it one rung at a time: unknown, then discovered, then crawled, then indexed. Knowing which rung a page sits on tells you what to fix. Rewriting a page stuck at “unknown” is wasted work, because nobody has read it.

Search Console saysWhat it meansWhat actually fixes it
URL is unknown to GoogleNo record of the URL at allGet it into a sitemap Google rereads, and link to it internally
Discovered, currently not indexedKnown, never fetchedCrawl priority: internal links, sitemap freshness, authority
Crawled, currently not indexedFetched, judged, declinedRewrite the page. It lost on merit
IndexedEligible to rankNothing here. Any problem is a ranking problem

The middle two decide where your effort goes, and they are the pair most often collapsed into “not indexed.” Rewriting a discovered-not-indexed page changes nothing, because Google never read the old version.

Why does Google deindex pages?

Google deindexes pages for six reasons, and in my experience they are wildly unequal in how often they turn out to be the answer.

Something on the page told it to leave. A noindex meta tag, an X-Robots-Tag header, or a canonical tag pointing somewhere else. This is the most common cause and the most embarrassing one, because it is nearly always shipped by accident during a redesign or a plugin change. A robots.txt rule does its own kind of damage: Google stops reading the page, and an already-indexed URL can linger as “Indexed, though blocked by robots.txt” with no description and nothing to rank on.

A canonical pointing at a redirect. This is the version nobody checks, so here is a real one. Every page on this site declared a canonical URL ending in .html, while the host serves the extensionless URL and 308-redirects the .html one. So every page was telling Google the real version of me is over there, and pointing at a redirect. Search Console confirmed the damage: googleCanonical did not match userCanonical on every page inspected. Google had overridden the declared canonical sitewide. One line in a layout file, wrong on all 90 URLs. The fix shipped on 2026-08-07, and three months later Google still held the old .html canonical on three URLs it had not recrawled since. A canonical fix only lands when Google next visits each page.

The page started returning errors. Google’s documentation on HTTP errors is blunt about it: a previously indexed URL that returns a 404 or 410 is removed from the index, and one that keeps returning a 5xx server error is held for a while and then dropped. A deploy that breaks a single template can take out a whole page type this way. Pages that load empty or show an error while returning a 200 get flagged as soft 404s.

A removal request, possibly one you never filed. Search Console’s Removals tool hides a URL from results for about six months. It also lists requests made by anyone through Google’s public outdated-content tool, and in July 2025 Search Engine Journal reported a bug that let strangers remove live pages by submitting a capitalized variant of the URL that returned a 404. One publisher lost more than 400 articles before Google fixed it.

A manual action. Rare, and unambiguous: Search Console tells you, in writing, under Security & Manual Actions. If there is no message there, you do not have a penalty, and a reconsideration request will do nothing. Plenty of advice recommends filing one anyway. Do not. When there is a real manual action, Google says reconsideration reviews can take several days or weeks, and fixing only some of the listed pages earns no partial return.

Crawl demand collapsed. This one is neglect. Google crawls what it thinks is worth crawling, and a young or low-authority domain gets a small allowance that goes to the pages already proven to earn clicks.

That last one is worth measuring before you assume it. This site’s Crawl stats report (Settings, then Crawl stats) logged 1,516 crawl requests over the 90 days to August 2026 at a 263 ms average response time. The standard advice for Google not crawling website pages often enough is to make the site faster. At 263 ms, speed is not the constraint, and no amount of optimization would buy another crawl. Google crawls this site exactly as much as it wants to.

Crawl capacity and crawl demand are different problems. Fixing your server speed when the issue is demand is like repaving a road nobody wants to drive down.

Why is Google not indexing all my pages?

Google is not indexing all your pages because it has decided some of them are not worth a crawl yet, and on a young or fast-growing site that is the usual reason. Rule out the technical causes above first. When they come back clean, what is left is demand.

This site is the worked example. On 2026-11-16 I ran URL Inspection on every sitemap URL that had earned zero impressions in the previous 28 days.

The technical side was clean on every one of them. Each returned a 200, none carried a noindex, every canonical pointed at itself, and Google had read the sitemap on 2026-11-09 with zero errors. Nothing was blocking Google, and 30 pages were missing from the index anyway.

Two things explained it. The site had been publishing faster than Google was willing to crawl it, so the newest pages (service pages and fresh blog posts) waited at the back of the queue as “unknown” or “discovered.” And those new pages had thin internal links: few links from the pages Google already crawls often, so nothing signalled that they mattered. The same pattern showed up in August, when 19 posts in one new cluster went live, most of them within about ten days, and every one spot-checked sat at “Discovered” with zero impressions, while the older clusters indexed normally.

The eight crawled-not-indexed URLs were a different case. Five were older posts Google had last fetched between July and September, all rewritten since. Google’s verdict on those five is a verdict on versions of the pages that no longer exist, so what they need is a recrawl. The other three were new service pages Google read on 2026-10-28 and declined.

If your Page indexing report looks like this, a technical SEO audit will come back clean and tell you nothing. The levers are internal links from pages Google already visits, a sitemap with real lastmod dates, and a publishing pace the crawl can keep up with.

How to check whether you are deindexed

Check one URL with the site: operator, and check the whole site with the URL Inspection API.

The manual version takes ten seconds. Search site:yourdomain.com/the-exact-url. A result means indexed. No result is a strong hint, and Search Console’s URL Inspection tool settles it. Paste the URL there and it returns one of three verdicts: URL is on Google, URL is on Google, but has issues, or URL is not on Google, with the reason underneath (for example, “Page is not indexed: URL is unknown to Google”).

That verdict is the thing to trust. The Page indexing report (Indexing, then Pages) groups everything under “Not indexed” and buries the reason a click deeper, so a rising count of pages not indexed tells you almost nothing on its own. Some of those reasons are your own instruction working correctly, like “Excluded by ‘noindex’ tag” on a thank-you page. Read the reason attached to the URL.

That works fine for one page and falls apart for a catalog. Clicking through 90 URLs by hand is an afternoon, and it is the reason most sites never find out how bad their coverage is. The URL Inspection API returns the same verdict programmatically, at up to 2,000 inspections per property per day, so you can sweep every URL in your sitemap and get a full picture. That sweep is where the 57-out-of-90 number above came from, and it surfaced two things clicking around never would have: two case-study pages that had never been crawled at all, and a whole 19-post cluster with zero impressions in 90 days.

One more check most people skip. In Search Console, open Indexing, then Sitemaps, and compare the “Last read” date against your last publish. Google held a copy of this site’s sitemap that was two days stale and 11 URLs short, which meant a batch of new posts did not exist as far as Google was concerned. The sitemap also carried no lastmod dates at all, so nothing signalled that a sitewide fix had shipped. If your sitemap has no lastmod, you have thrown away the one freshness lever you fully control.

Google not indexing sitemap URLs is usually this. The file was read weeks ago, the new URLs were not in it then, and resubmitting forces a fresh read. Three dead sitemaps from a previous platform were still registered on this property at the same time, all returning 404, quietly padding the submitted count with URLs that no longer existed.

Two last checks take a minute each. Open Security & Manual Actions, then Manual actions: an empty report means no penalty. Then open Indexing, then Removals, and read the outdated-content requests for any nobody on your team made.

How to get reindexed

Fix the cause first, then request indexing. In that order, because requesting a crawl of a page that still says noindex just gets you a faster no.

Two things fall out of that. Google will discard a URL it already knew about, so “Discovered” is not a safe resting place. And a manual request genuinely moves a page, fast, when the sitemap alone had not moved it in three weeks.

The sitemap mattered too, once Google actually reread it. Google read the resubmitted file on 2026-08-09, and by the next day the count of URLs Google had no record of fell from 15 to 3. The redesign post that went from unknown to crawled finished the climb: unknown, crawled, indexed, in under 48 hours.

The catch is throughput. Google says only that there is a daily limit on indexing requests, and in my use that limit lands at roughly 10 to 12 URLs per property per day, which makes Request Indexing a scalpel. You cannot recover a 500-page site with it, and if you find yourself wanting to, the real problem is that Google is not crawling you on its own. Google’s URL Inspection help also says a request does not guarantee inclusion, and that indexing typically takes a day or so and can take a week or two.

The durable fixes are less satisfying and they work on the whole catalog at once:

  1. Remove the block. Strip the noindex, fix the canonical, unblock the path in robots.txt, and make sure the URL returns a 200. Confirm on a recrawled page that googleCanonical and userCanonical finally match.
  2. Resubmit the sitemap so Google rereads it, and ship real lastmod dates so the reread carries a freshness signal. Google’s own guidance recommends a sitemap with lastmod when many pages are new or updated.
  3. Give the page internal links. A page nothing links to is a page Google has little reason to prioritize. Internal linking is the cheapest crawl signal you own. On client sites where indexing trouble traced back to weak internal linking, I add the hub to the site footer and ship a related-pages block that links every orphaned page, with each link pointing straight at the final URL and never through a redirect.
  4. Clear the crawl path of junk. On a cybersecurity SaaS account, I added targeted robots rules to keep parameter, staging, and feed URLs out of the crawl path during a redirect migration, and roughly 280 URLs cleared from the crawled-not-indexed backlog. That is a coverage cleanup: the point is that the crawl budget stops being spent on URLs you never wanted in the index. The full robots.txt playbook covers where the line sits.
  5. Give Google a reason to want the page. For a page Google crawled and declined, the fix is the page itself. Crawled but not indexed is a verdict: Google found the page, evaluated it, and decided it was not worth showing. Add the thing only you have, a first-party number, a worked example, an answer the indexed pages do not give, and then request the recrawl.

How to deindex a page on purpose

Deindex a page on purpose with a noindex rule Google can read, or by removing the page so it returns a 404 or 410. Google’s guide to blocking indexing covers the mechanics, and every switch on this list is also a way sites deindex themselves by accident.

  • A noindex meta tag or X-Robots-Tag header. The page drops out on Google’s next crawl. The header version works for PDFs and other non-HTML files. The page has to stay crawlable: if robots.txt blocks it, Google never sees the noindex.
  • A 404 or 410. Delete the page, and Google removes the URL once it recrawls and gets the error.
  • A password. Content behind a login cannot be indexed.
  • The Removals tool. It hides a URL from results for about six months, and Google keeps crawling the page the whole time. Use it for something urgent, then put a noindex or a 404 in place before the six months run out.
  • robots.txt is the wrong tool for removal. A disallowed URL that other sites link to can still appear in results, and Google does not support a noindex rule inside robots.txt.

Read that list again with a redesign in mind. A staging noindex that ships to production, a robots rule copied from the old site, a template that starts returning 500s: each one is a deliberate deindexing tool, used by accident. A site audit tool crawl after each deploy flags all three.

When it is not deindexing at all

If your pages are still in the index, the fix lives somewhere else.

Confirm it first, because it takes one search. Run site: on a URL you think you lost. If it returns, you are indexed, and what you are looking at is a ranking problem or a click problem wearing a deindexing costume.

A website indexed but not showing for the query you care about is the most common version of this, and it is a ranking problem with a completely different fix. The page is eligible; it is losing to something else. On a legal SaaS platform set, nearly every target BOFU keyword reached the top 10 once it went into the page’s H1 and the benefit copy moved down into supporting H2s.

This site has 23 pages in exactly that position. On 2026-11-19 I re-inspected every sitemap URL that earned zero impressions between 2026-10-19 and 2026-11-15 and sat outside the not-indexed groups above. All 23 came back “Submitted and indexed,” with Google’s canonical matching the declared one: two category hubs, a run of GEO and SEO-services posts, a calculator, and a case study. Several had last been crawled in August or September. Those pages are in the index and earned zero impressions in four weeks. Being indexed only makes a page eligible to appear, so the work on these pages is demand and rankings: better targeting, stronger internal links, and more authority behind them.

A sharp sitewide impressions drop in mid-September 2025, paired with a better average position, was a reporting change. Google stopped honoring the num=100 results parameter that rank-tracking tools relied on, and Search Console impressions fell across many sites as tool-generated impressions dropped out of the data.

Impressions holding while clicks fall is the great decoupling, and it happens on pages that are fully indexed. A sharp drop dated to a known update is a core update question. Rankings sliding with no change in index status is why your rankings dropped, and a broader traffic slide is why your organic traffic is dropping. If the pages were never indexed in the first place, you have a website not showing up on Google at all, and that diagnosis starts earlier.

FAQ

What does indexing mean in simple terms?

Indexing means Google has stored a copy of your page in the database it pulls search results from. Crawling is Google fetching the page; indexing is Google deciding to keep it. A page has to be indexed before it can rank for anything, which is why a deindexed page loses every ranking at once.

What does “non-indexed” mean?

Non-indexed means Google is not storing the page and will not show it in results. It covers three of the four states above: unknown to Google, discovered but not crawled, and crawled but not indexed. Search Console groups them under “Not indexed” in the Page indexing report, which is why so many teams treat one problem as three, or three problems as one. Open the report and read which specific reason is attached to each URL before you fix anything.

How does de-indexing work?

Google removes the URL from its index, either because the page told it to or because Google decided it no longer merits storage. Removal by instruction happens on the next crawl: a noindex tag, an X-Robots-Tag header, or a 404. A Removals request is a different mechanism: it hides the URL from results for about six months while the page stays crawlable. Removal by decision is quieter and has no notification. The page simply stops appearing, and the first symptom is usually impressions going to zero in Search Console.

Should I let search engines index my website?

Yes, for every page you want people to find through search. Keep a noindex on pages with no search value: thank-you and confirmation pages, internal search results, staging and test environments, and thin tag or filter pages. A workable rule for a B2B SaaS site: if a buyer would never land on the page from Google on purpose, it does not need to be indexed.

What is the difference between discovered not indexed and crawled not indexed?

Discovered means Google knows the URL exists but has never fetched it. Crawled means Google fetched it, read the content, and chose not to index it. The distinction decides where your effort goes: discovered is a crawl-priority problem, so you work on internal links, sitemap freshness, and giving Google a reason to spend a crawl. Crawled is a quality verdict, so rewriting the page is the only thing that helps. Fixing a “discovered” page by rewriting it changes nothing, because nobody has read the old version.

What does status not indexed mean in Search Console?

It means the URL is not eligible to appear in search results, and the specific reason sits next to it in the Page indexing report. Some reasons are your instruction and completely fine, like “Excluded by ‘noindex’ tag” on a thank-you page or “Alternate page with proper canonical tag” on a duplicate. Others are problems, like “Discovered - currently not indexed” across a whole cluster. Read the reason. A rising “not indexed” count is not automatically bad news.

How long does it take to get reindexed?

For a manually submitted URL with the underlying cause already fixed, hours to a few days. On this site, one submitted URL went from unknown to crawled within hours and was indexed the next day, under 48 hours end to end. Google’s own guidance says a day or so is typical and a week or two is possible. Recovery without a manual request runs on Google’s schedule and depends on how often it crawls you at all: at 1,516 crawl requests over 90 days, a page nothing links to can wait weeks. Fix the cause, request the important URLs, then let internal links and sitemap freshness carry the rest.

Can a deindexed page rank again after it comes back?

Yes, and usually at a similar position, provided the deindexing was technical and short. The page keeps its links and its history. What does not survive well is a long absence combined with a quality cause: if the page was dropped because Google judged it thin, reindexing it unchanged puts the same page in front of the same bar. Fix the underlying reason, or you will be back here in a month.