How a UK SME Should Read the Page Indexing Report Before Expecting an AI Citation

Answer first: If the site has fewer than 500 pages, Google’s non-expert guide says you probably do not need the Page indexing report. Search for the service URL first. Open the report when those searches come back empty, or when you need the reason a group of URLs was left out. A reason in the “not indexed” table means Google has not indexed that URL, which fails the supporting-link test in its AI features guide. One robots.txt case, covered below, is the exception the Page indexing help page prints: the page might still be indexed from other information, without a fetch. A URL marked indexed has passed the index gate and nothing more. The count is not a citation, a click, or an enquiry. Recheck the help page before you brief anyone from a row. This article records it for a reader on 19 October 2026.

When should a small firm open this report?

Google’s Page indexing report help page splits the audience. For someone new to indexing, or for a small site, it says: if you have fewer than 500 pages, you probably do not need this report. Use Google itself.

• A site: search on your host shows a sample of pages Google knows about.
• The same search plus words from the page shows whether that subject is in the sample.
• A search for the exact URL shows whether that address is indexed.

Google says to open the report only if those searches return nothing for the site, and only if you are willing to read it properly. Skipping the reason for each row wastes the afternoon.

The same page says this report is not how you investigate one URL. That job is the URL Inspection tool, already written up in the inspection guide. You also cannot search or filter this report by URL. If you care about /services/boiler-repair/, inspect that address. Do not hunt for it in a sample table and treat a miss as a verdict.

Two counts sit on the summary. The indexed total and the not-indexed total above the chart are what Google calls complete and accurate from its side. Small differences against your own page count can still happen. The example list under a reason is capped at 1,000 rows, and Google says it is not guaranteed to include every URL in that status, even when the status holds fewer than 1,000. A service page missing from the sample can still be inside the total. Inspection is the check.

The report covers the property you have open. A staging host is not a reading of the public site. If customers use two hosts, open both properties. How a .co.uk property sits beside a .com property is already in the control guide.

What does “indexed” mean on this report?

Google’s definition on the help page is short. Indexing means Google found the page, processed it, and placed it in the Google index, where it may be eligible to appear in Google Search and in other Google services, such as Discover. “May be eligible” is the strength of that sentence. Indexed is not “showing”, and it is not “showing for the query you typed”.

The help page also says an indexed page is not guaranteed to appear in your own results. Results vary with history, location, and other factors. If Search Console says the URL is indexed and your search does not show it, Google’s instruction is to assume it is indexed and eligible to appear. Your browser is the wrong argument against the report.

The other half of the summary is the “Why pages aren’t indexed” table. Click a row for the URLs Google grouped under that reason, and for the history of that reason. A second table, “Improve page experience”, lists issues that did not stop indexing. Google says those are worth fixing because they make the page harder to understand. They are not the same thing as a URL left out of the index. Read the table name before you treat a warning as a removal.

Each not-indexed row has a source. Google’s help page says the source is either Google or the website, and that you can generally fix only the rows whose source is Website. Read the cell. Do not assign a source from memory, or from this article. The page recorded here does not pin a source onto each reason.

Not indexed is often the correct outcome. Google says it does not index pages blocked by robots.txt or by noindex, duplicates of other pages, or pages that are a poor fit to index, and its example is filter variations of the same collection. The goal it names is narrower: the canonical version of each important page. A row marked duplicate or alternate is usually a good sign on this page, because it means Google found the canonical and indexed that one. Confirm which URL that is in URL Inspection, rather than assuming the service page won.

Which reasons should change what I do next?

The help page lists many reasons. These are the ones that change the next action for a small firm. The labels below are the labels on that page.

URL marked ‘noindex’. Google fetched the page, saw a noindex directive, and did not index it. If you wanted the page out, the row is finished. If you wanted it in, remove the directive. The help page tells you to confirm it in URL Inspection under Indexing allowed, including on the live test. If the live test no longer sees noindex, you can request indexing. If it still sees the tag, requesting indexing will not overwrite it. A theme or a plugin can add the tag. The inspection guide is how to read the HTML Google received. The control guide is where noindex sits next to nosnippet.

URL blocked by robots.txt. Google says this does not guarantee the page will stay out of the index. If it can learn about the URL without loading it, there is a very small chance the page might still be indexed. To keep a page out for certain, Google’s instruction is to remove the robots.txt block and use noindex. Blocking the file also hides any noindex you thought was on the page. Which token is Googlebot, as distinct from Google-Extended, is the citations guide.

Indexed, though blocked by robots.txt. This one sits in the warnings, so it did not stop indexing. Google says it respects robots.txt and will not request the page, but it can still index the URL from information on a page that links to it. Any snippet in Search results will probably be very limited. Do not brief this warning as “the service page is indexed, so the answer on it can be quoted.” Google is describing an index entry built without a fetch of that URL.

Page indexed without content. Also a warning. The URL is in the index, and Google could not read the content. The reasons it offers are that the page might be cloaked, or that it might be in a format Google cannot index. It says this is not a robots.txt block. Inspect the URL and read the coverage details. The AI features guide says important content should be available as text. This warning is the stronger reading: Google says it could not read the content at all.

Server error (5xx). The server returned a 500-level response. Google’s troubleshooting note says Googlebot could not access the URL, the request timed out, or the site was busy, so the fetch was abandoned. It tells you to check the host-status verdict in the Crawl Stats report. The same note says server errors can be transient: a live test can succeed later, because the later fetch is not the failed crawl. One 5xx row is not a reason to buy a citation product.

Soft 404. Google thinks the page is a friendly “not found” that did not use a 404 status. It recommends a real 404 for pages that are truly gone, and more information on the page so Google can tell it is not a soft 404. The check it names is a live URL inspection and the screenshot under View tested page. A service URL that renders an empty not-found state while returning success belongs in this reading. Do not “fix” it by publishing a copy of the page for every village.

Not found (404). The URL returned 404. Google may have found it from a link, or it may be an old address. Google says it will probably keep trying for a while, that you cannot tell it to forget a URL permanently, and that it will crawl the URL less often over time. A 404 is not automatically a problem when the page was removed and nothing replaced it. If the page moved, Google says to use a 301 to the new address. The report shows URLs that returned 404 in the past month, not a lifetime archive. In the section on removed pages, Google also says the only case in which a URL will not be crawled is when it returns noindex.

Blocked due to unauthorized request (401), access forbidden (403), or another 4xx. A 401 asked for authorisation. Google says to check in a private window. A 403 means credentials were offered and refused, but Googlebot does not offer credentials, so the server is refusing a request it should be able to serve. The page will not be indexed while that stands. A quote form that sits behind a login is invisible to this crawl. The form itself, for a person and for a browser agent, is the quote form guide.

What does “Crawled - currently not indexed” mean?

Google’s entire description is this: the page was crawled and not indexed. It may or may not be indexed in the future. There is no need to resubmit the URL for crawling.

That is the whole instruction. The help page does not name a word count, a score, or a content type for this row. Do not paste a cause onto it from a forum post, and do not paste one from this article. Resubmitting the sitemap, or clicking Request indexing because a chat did not mention you, is not the step this reason prints.

Use URL Inspection when you need the current state of that one URL. The page indexing report can lag a fix. Google says a URL can still show an issue you have already corrected, because the correction came after the last crawl. Read the crawl date before you treat the row as today’s page. The same help page says the live test does not check every condition this report covers. Duplicate and canonical decisions are the ones it names as absent from the live test. A clean live test does not erase this row.

If you later change the page in a substantial way, a different section of the same help page applies. Google says it recrawls according to how often it thinks a page changes, and that a quiet site on a slower refresh is fine when nothing changed. If the page did change substantially and Google has not noticed for a week or more, you can ask for a recrawl. It also says not to bother asking unless the change matters. The button, and the ranges printed on the URL Inspection help page, are in the inspection guide. Do not average those ranges with the ones on the Page indexing page. They are different pages.

What you should not do is invent a length target. Google’s optimisation guide says there is no ideal page length, and that pages are for the audience. It also says that creating a separate page for every variation of a query, mainly to manipulate rankings or generative AI responses, violates its scaled content abuse policy. A stack of near-copy town pages is not a treatment this status describes. One page that states the service, and the places you actually cover, is the page to inspect. How to write the coverage sentence is the local citations guide. How to shape the page around one question is the service-page guide.

What does “Discovered - currently not indexed” mean?

Google says the URL was found and has not been crawled yet. Typically, Google wanted to crawl it but expected that crawl to overload the site, so it rescheduled. That is why the last crawl date on the report is empty.

This is not a noindex row, and it is not a robots.txt row. Nothing on this description says the page was rejected after a fetch. There was no fetch to reject. Do not “fix” an empty crawl date by publishing another URL.

The help page’s discovery notes sit beside this status. Google learns a URL from a link on a known page, from a sitemap, or from a crawl request. For a new site, it says the best first step is to request indexing of the homepage, which should start a crawl, and that navigation from that homepage should be enough for it to find the other pages if the navigation is real. A new page or a new site can take a week or so before crawling and indexing start. After a URL is known, it can take up to a few weeks before Google crawls some or all of the site. Indexing is never instant, including when you submit a crawl request. Google also says it does not guarantee that every page enters the index. Those are four separate sentences. They are not a delivery date.

A host that publishes pages for you may tell Google itself. Google’s examples are Wix and SquareSpace. Check that host’s own documentation. Do not assume a template has done it.

The sitemap dropdown is a filter, not a reason. The default is all known pages. You can limit the view to URLs in a sitemap submitted through the Sitemaps report or named in robots.txt, to URLs that were not in such a sitemap, or to one sitemap. Google says a URL still counts as submitted if a sitemap lists it, even when a link also revealed it. Unsubmitted is not one of the not-indexed reasons above. A page can be indexed without a sitemap. Submitting a sitemap does not move a URL out of “Discovered - currently not indexed” on a clock this page prints.

When is a duplicate, a redirect, or a missing page the right row?

Duplicate without user-selected canonical. The page duplicates another, and it does not name a preferred canonical. Google chose the other URL and will not serve this one. The help page says this is not an error. Inspect the URL to see the chosen canonical. If you think the choice is wrong, mark the canonical yourself. If you think the pages are not duplicates, Google says the content has to differ substantially. Swapping a town name in the title is a weak claim that they differ.

Duplicate, Google chose different canonical than user. You named one canonical. Google indexed a different URL. Inspect the address and read both the Google-selected canonical and the one you declared. The help page says Google will not select your declared canonical if that page is not similar to the one it crawled. Similarity is what “duplicate” means on that page.

Alternate page with proper canonical tag. The URL points at a canonical that is indexed. Google says there is nothing to do. It also says Search Console does not detect alternate language pages. A Welsh page, or a second-language URL, will not appear under this reason. A missing row is not proof that the language page is indexed, and it is not proof that it failed.

Page with redirect. This URL redirects, so this URL will not be indexed. The target might be indexed, or it might not. Inspect the target before you judge the old address. Google says the indexed reading describes the URL you typed and ignores the redirect, while the live test follows the redirect and does not tell you that it did. A firm that moved /contact/ to /quote/ can look at a green live test and think it checked the old URL. The inspection guide walks through that screen. This article only says: the redirect row is about the address in the table, not about the page it now sends people to.

A drop in the indexed count, with no new error, is also a documented pattern. Google says you might have added a robots.txt block, a noindex, or a login. Look for a rise in a not-indexed reason that lines up with the drop. More not-indexed URLs than indexed ones can be normal on this page, for the same reasons: a broad robots.txt rule, or a large set of sorted and filtered duplicates that should not be indexed. On a ten-page firm, that pattern is a reason to read the rows. It is not, on this help page, named as a penalty.

Does an indexed count mean an AI citation?

No. Google’s AI features guide says that to be eligible as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet, and it must meet the ordinary technical requirements. There are no extra technical requirements, and no special schema. Meeting them still does not mean Google will crawl, index, or serve the page.

The optimisation guide adds a second gate. Besides those technical requirements, the site has to be included in Search generative AI features in Search Console. An indexed URL on an excluded property is still indexed. Exclusion does not rewrite this report into “not indexed”. Google’s Search generative AI control says exclusion keeps your links and content out of AI Overviews, AI Mode, and generative AI features in Discover, and that the control is not a ranking signal for the rest of Search. The control guide is that switch. This report will not show it.

Snippet eligibility is the other half of the AI features sentence, and this report’s indexed status does not list nosnippet. The reasons recorded here do not include it. A page can be indexed and still carry a preview limit. Those tags, and the fact that nosnippet also removes the ordinary text snippet, are in the control guide. Google’s AI features page also says AI Overviews are shown when its systems decide they add something beyond ordinary results, and that they often do not trigger. A query with no overview is not a failed row in this report.

Impressions inside AI Overviews and AI Mode, once a link is actually shown, are the Generative AI performance report. The optimisation guide names that report as the measurement, and it says no third-party tool has access to Google’s internal ranking or AI systems. A chat screenshot is not a substitute for either report.

ChatGPT, Perplexity, Bing, Copilot, and Apple do not read this screen. Their controls are the citations guide and the Perplexity, Bing, Copilot and Apple guide. For Google Search, the optimisation guide says answer-engine optimisation and generative-engine optimisation are still SEO. An llms.txt file is not how a URL leaves the not-indexed table. Google says Search ignores that file.

What should I check on the day I open it?

Do this on the property customers use. Write the date next to what you find.

1. If the site has fewer than 500 pages, search for the exact service URL, and run a site: search on the host, before you open the report. Google’s non-expert guide says that is the first check.
2. If you do open it, read the indexed total and the not-indexed total above the chart. Then read the reason rows, not the shape of the chart.
3. For each reason that contains a URL you care about, read the source cell and the help text for that reason. Fix Website-sourced blocks you did not mean: a stray noindex, a robots.txt disallow, a 401 or 403, a soft 404, a redirect to the wrong place.
4. Do not resubmit a URL whose reason is “Crawled - currently not indexed”. Inspect it instead, and read the crawl date.
5. For “Discovered - currently not indexed”, check that a normal link exists from a page Google already knows, usually the homepage or the service list. Do not treat an empty crawl date as a broken page.
6. If the row is a duplicate or an alternate, inspect the Google-selected canonical and decide whether that is the URL you want understood.
7. If the indexed HTML is available in URL Inspection, confirm the answer is text. This report does not show that HTML.
8. If the question is an AI Overview or AI Mode impression, leave this report. Read the include setting, then the performance report.

What should I not claim?

• That “indexed” is an AI Overview, a citation, a click, or an enquiry.
• That “Crawled - currently not indexed” means a score, a word count, or a penalty. The help page does not say that. It says there is no need to resubmit.
• That “Discovered - currently not indexed” means you blocked the page. Google describes a crawl it rescheduled.
• That every URL should be indexed, or that the example table is a complete export. Google caps the examples at 1,000 and says the list can be incomplete.
• That a sitemap submission, an llms.txt file, or FAQ markup moves a URL into AI Overviews. The AI features guide says there is no special file or schema for those features.
• That this report measures ChatGPT, Perplexity, Bing, Copilot, or Apple.
• A result, a town, or a before-and-after Halo has not published with evidence. This guide does not add one.

Where does Halo fit?

Halo’s search, AEO and content service covers the public page: the answer in the first screen, internal links, and structured data that matches the visible words. The Page indexing report sits in the Search Console property you verify. Halo does not need that login to read the URL. If you already have the screen, the useful note is the reason, the source, and the crawl date for the one address you care about.

To start, send the URL through the free AI audit. Halo does not guarantee indexing, impressions, citations, or enquiries.

FAQ

Does an indexed page in this report appear in AI Overviews?

No. Google’s AI features guide says a supporting link has to be indexed and eligible for a snippet, and that indexing and serving are not guaranteed. The optimisation guide adds that the property has to be included in Search generative AI features. This report’s indexed count is only the index half of that gate. It is not an impression or an enquiry.

Should I resubmit a URL marked crawled but not indexed?

Google’s page indexing help says the page was crawled and not indexed, that it may or may not be indexed later, and that there is no need to resubmit it for crawling. Inspect the URL if you need the current reading. Ask for a recrawl only after a substantial change Google has not picked up, which is a different section of the same help page.

Does “Discovered - currently not indexed” mean I blocked the page?

No. The help page says the URL was found and not crawled yet, typically because crawling it was expected to overload the site, so the crawl was rescheduled. That is why the last crawl date is empty. It is not the robots.txt row, and it is not the noindex row.

If my site has fewer than 500 pages, do I still open this report?

Google’s non-expert guide says you probably do not. Search for the important URLs first. Open the report if those searches return nothing and you are willing to read each reason. A small site can still have one service page that is not indexed. URL Inspection is how you read that single address.

Will this report show whether ChatGPT cited the page?

No. The report is Google’s index of URLs it knows about on the Search Console property you opened. ChatGPT, Perplexity, Bing, Copilot, and Apple are not rows in it. A pass or a fail here does not travel to those products.

Does an Isle of Man firm get a different Page indexing report?

No. The help page does not describe an Isle of Man report or a UK-only report. You get the report for each property you verified. A .im host and a .co.uk host are two reports if they were added as two properties. The page still has to name where you actually work.

Sources

• Google Search Console Help, Page indexing report, including the fewer-than-500-pages guidance, the site: checks, the 1,000-row example limit, the indexed and not-indexed totals, the source column, the reason definitions used above, the sitemap filter, and the notes on new pages and recrawl requests.
• Google Search Console Help, URL Inspection tool, the one-URL check this report tells you to use instead of searching the example table.
• Google Search Central, AI features and your website, including supporting-link eligibility, the statement that indexing and serving are not guaranteed, and the note that AI Overviews often do not trigger.
• Google Search Central, Optimizing your website for generative AI features on Google Search, including the Search Console include requirement, scaled content abuse, the statement that there is no ideal page length, the Generative AI performance report, and the statement that Google Search ignores llms.txt.
• Google Search Console Help, Search generative AI control, including Exclude, and the statement that the control is not a ranking signal for the rest of Search.