How to Fix "Crawled - Currently Not Indexed" in Google Search Console (2026 Guide)

You published the page. Googlebot crawled it. The URL still is not in Google's index.

That is what "Crawled - currently not indexed" reports, and the maddening part is that the status does not tell you why. The label covers several different situations, and the standard advice (improve your content, add links, click Request Indexing) addresses only some of them.

This guide separates them. You get a five-gate diagnostic framework created for this article that pinpoints where a URL is most likely failing, a triage system that sorts thousands of URLs in under an hour, a differential-diagnosis table that maps symptom patterns to likely causes, and honest observation windows including the cases that never recover.

It is built on Google's official documentation, public statements from Google's Search Relations team through July 2026, third-party commercial indexing research covering roughly 1.7 million monitored URLs on 18 customer websites, and recurring patterns reported by site owners on Reddit, Quora, and the Shopify and Google Search Central forums.

The status by itself does not reveal the cause for an individual URL. Technical accessibility, rendering, canonicalization, duplication, soft-404 signals, sitewide quality patterns, and page-level usefulness can all contribute. Google representatives have said quality is sometimes a reason, but the Search Console status is not a diagnosis.

To investigate it:

1. Filter the noise. Feed URLs, pagination, tag archives, parameter and filter URLs commonly appear in this report. Strip them out before diagnosing anything.
2. Establish the URL's history. In Search Console → Performance, set the range to 16 months and filter to the exact URL. Positive impressions confirm the URL appeared in Search during that period. Zero impressions do not prove it was never indexed; an indexed page can receive no impressions. Use URL Inspection and the Page indexing report as your primary indexing checks.
3. Run a live render test (URL Inspection → Test Live URL → View Tested Page). If your main content is missing from the render, you likely have a technical fault to fix first.
4. Rule out canonical conflicts. Compare the user-declared canonical against the Google-selected canonical.
5. If the render is clean and the URL is unique, page-level usefulness and redundancy become plausible contributing factors. They remain hypotheses, not confirmed causes.
6. Add relevant, crawlable internal links from established indexed pages.
7. Prune or consolidate weak pages sitewide. Google has said sitewide quality concerns can reduce how much of a site it crawls and indexes.
8. Request indexing once, after you have materially changed the page. Resubmitting an unchanged URL queues a recrawl, not a re-evaluation.

Observation windows (author's suggested checkpoints, not Google guidance): technical fixes often show movement within a few weeks. Duplicate consolidation commonly takes longer. Usefulness-driven cases can take several months, and a meaningful share never recover. Indexing may occur quickly, take several weeks, or not occur at all if Google continues to exclude the URL.

What's normal: every site has unindexed URLs, and Google's Page indexing documentation states you should not expect 100% of your URLs to be indexed. Judge the report by which pages are in it, not by the number. Ten unindexed money pages is urgent. Ten thousand unindexed /feed/ URLs usually is not.

What "Crawled – Currently Not Indexed" Actually Means

Google's definition, from the Page indexing report documentation linked above, is deliberately non-committal:

"The page was crawled by Google but not indexed. It may or may not be indexed in the future; no need to resubmit this URL for crawling."

Three things are worth pulling out of that sentence.

"Crawled" means Googlebot fetched the URL. It does not confirm that Google successfully loaded every resource, completed rendering, or evaluated the complete page. Rendering is a separate step that can be delayed or fail, and Google's JavaScript SEO documentation is explicit that important content should be present in the rendered HTML and accessible to Googlebot rather than assumed. The status confirms Google saw something. It does not confirm Google saw what your visitors see. That distinction is one of the most valuable diagnostics in this guide.

"Not indexed" describes a current state, not a permanent one. The crawl completed. At the end of the pipeline, the URL was not added to the index at this time.

"No need to resubmit" is Google stating that the resubmit button is not the lever most people assume it is.

Where your page may have failed in the pipeline

Every URL travels through several stages:

Discovery → Crawling → Rendering → Indexing → Ranking
              ↑                        ↑
        "Discovered –            "Crawled –
     currently not indexed"   currently not indexed"
        reported here            reported here

A URL with this status was discovered and crawled. Whether rendering completed successfully is something you have to verify separately, not something the status confirms. What is certain is that the URL is not in the index right now, and an unindexed page cannot rank or receive impressions in Google Search.

The question worth asking

Indexing is not free. Google stores, refreshes, and serves from a finite index, and the evaluation of a given URL is comparative rather than absolute.

In the July 2026 Search Off the Record episode on reading the indexing report, John Mueller described this directly: when a great deal of comparable material already exists, he questioned what value a further version of the same coverage adds to the index. He framed the consideration as what the index gains by including a page, rather than simply whether the page is good.

That framing explains why many site owners are baffled. They run their page against every quality checklist, it passes, and it still does not get indexed. Being comparable to what already ranks may not be sufficient on its own.

Treat this as one plausible contributing factor among several, not as a confirmed reading of your specific URL.

Is it an error? Is it a penalty?

• It is not an error. Google classifies it as a valid, non-error state. The crawl completed.
• It is not, by itself, evidence of a manual action or penalty. Manual actions appear in Search Console under Security & Manual Actions. This status alone does not prove that no spam, security, or manual-action problem exists elsewhere on the site. If you are seeing broader symptoms, check the Manual Actions and Security Issues reports directly.
• It does not directly harm your other pages. A large volume of affected URLs may correlate with sitewide quality perception, which is why mass occurrences deserve attention, but correlation here is not established causation for any individual URL.

The Reframe: Two Hidden Variants of One Status

Here is a finding most guides on this topic miss, and it changes how you read your report.

Indexing Insight, a commercial indexing-monitoring product that tracks status at scale through Google's URL Inspection API, reported in its analysis of the Page indexing report that across roughly 1.7 million monitored pages on 18 customer websites, 70% to 80% of URLs carrying the "crawled – currently not indexed" label had previously been indexed by Google.

Important scope caveats: this is third-party commercial research on a specific customer dataset, not a Google-wide benchmark. Those 18 sites may not represent all websites. Treat the finding as directional rather than universal, and do not assume the same proportion applies to your site.

Within that dataset, the affected URLs were largely not pages waiting in line. They were pages Google had indexed and later removed.

Search Console does not separate these two situations. Both display the identical status string, and they can call for different responses.

How to check the URL's Search history

1. Search Console → Performance → Search results
2. Set the date range to the maximum 16 months
3. Click + New → Page → URL exactly matches and paste the URL

Impressions greater than zero confirm the URL appeared in Google Search during that window, per Google's Performance report documentation. Tag it RECOVERY.

Zero impressions do not prove the URL was never indexed. An indexed page can sit in the index without receiving impressions, particularly if it never ranked well enough to be seen. Tag it UNCONFIRMED, and treat URL Inspection and the Page indexing report as your primary indexing checks rather than impression data.

At scale, pull the Search Analytics API for a 16-month window, aggregate impressions by page, and left-join against your filtered Search Console export on the full URL.

Framework #1 (author-created): The Deindexation Audit. Run the 16-month impression check across your entire affected list, not as spot checks. Then sort by peak historical impressions, descending. URLs at the top of that sorted list are your highest-value recovery candidates and are worth prioritising. Most people work this list in whatever order Search Console exported it, which is close to random. This is a prioritisation heuristic, not a Google-defined process.

For any RECOVERY page, pull the version from the Wayback Machine at the date it was last performing well and diff it against today. Content gets thinned by CMS migrations, plugin changes, template edits, and well-meaning "SEO refreshes" more often than teams expect.

Crawled vs. Discovered vs. URL Unknown to Google

These three statuses describe different points in Google's processing of your URL.

The key difference: Discovered means Google has not fetched the URL yet. Crawled means it has. They usually call for different investigations, and crawl-scheduling remedies applied to a "Crawled" URL often produce nothing, because the crawl already happened.

Crawl-budget management is mainly a concern for very large or rapidly changing sites. Google's large-site crawl-budget guidance describes sites with roughly one million or more unique pages with content that changes about weekly, or medium-sized sites of roughly 10,000 or more unique pages with daily-changing content. Google's November 2022 office-hours explanation covers the same distinction. Smaller sites are less likely to face crawl-budget constraints, but "less likely" is not "never," and server responsiveness still affects crawling at any size.

A costly misreading: people see "Discovered," assume a discovery problem, and build more sitemaps and more internal links. In Indexing Insight's customer dataset, 94% of URLs the URL Inspection tool reported as "URL is unknown to Google" were grouped under "Discovered – currently not indexed" in the Page indexing report. Within that dataset, many URLs in the Discovered bucket were not queued for crawling at all. Again, this is one commercial dataset, not a universal rule.

Rule of thumb: movement from unknown → discovered → crawled – currently not indexed → indexed is progress, even when the label still reads like a failure. A page moving from "Discovered" to "Crawled – currently not indexed" is a positive sign. Google went and looked.

If you have both statuses in volume, examine "Crawled" first, since sitewide factors can affect both.

Why This Status Exploded in 2025–2026

Getting indexed used to be closer to automatic. Several data points suggest it is more selective now. Read the scope column carefully; most of this is third-party research on limited datasets.

Two forces are frequently cited by practitioners:

1. Scaled content. Generative tools made producing competent text on any topic inexpensive. Google's spam policies address scaled content abuse, and enforcement of content policies can affect what enters the index.
2. Sitewide signals. Mueller has said that when Google's systems have strong concerns about a site's overall quality, they may crawl and index less across the property.

The practical takeaway, stated as the author's interpretation rather than Google policy: inclusion in the index is not automatic, and pages that do not clearly add something can sit in this status indefinitely.

The 5-Gate Index Eligibility Model

The following five-gate model is a practical troubleshooting framework created for this article. It organizes common indexing causes but is not an official Google model. Google has not published these five gates, and does not state that it evaluates URLs in this exact order. Use it as a checklist sequence, not as a description of Google's internal systems.

The logic is that the cheapest, most objectively testable causes should be eliminated before the expensive, subjective ones.

The critical proportion

Gates 1 to 4 are cheap, fast, and objectively testable. Gate 5 is expensive, slow, and subjective.

In the author's experience auditing affected URL sets, a minority fail at gates 1 to 4 and the majority do not resolve to a clear technical cause. No published figure quantifies this split across the web, so treat it as an operating expectation rather than a statistic. Run gates 1 to 4 regardless: they take minutes for a whole template, and finding a failure there saves months of content work.

It is also why many people who "fix" their technical setup see nothing change. There may have been no technical fault to fix.

The decision tree

START: A URL is stuck in "Crawled – currently not indexed"

│
├─ Is it a feed, tag archive, parameter, filter, pagination,
│  or other URL you never wanted indexed?
│     └─ YES → IGNORE. Consider noindex if you never want it eligible
│              (this recategorizes the URL in reports; it does not
│              remove it from reporting). Stop here.
│
├─ GATE 1: Does it return a stable HTTP 200?
│     └─ NO → Fix server/redirect/CDN issue. Stop here.
│
├─ GATE 2: Does URL Inspection say "Indexing allowed: Yes"?
│     └─ NO → Remove noindex (check HTTP headers too). Stop here.
│
├─ GATE 3: Does the live render screenshot show your main content?
│     └─ NO → Rendering fault. Unblock resources / server-render. Stop here.
│
├─ GATE 4: Does Google's selected canonical match yours?
│     └─ NO → Duplication. Consolidate and repoint internal links. Stop here.
│
├─ GATE 5a: Has it had impressions in the last 16 months?
│     ├─ YES → RECOVERY case. Diff against the Wayback version.
│     │         Something changed: your page, the competition, or the
│     │         signals Google is weighing.
│     └─ NO  → UNCONFIRMED. No Search history recorded; this does not
│               prove it was never indexed. Verify with URL Inspection.
│
└─ GATE 5b: Run the Delta Test. Score it out of 12.
      ├─ 10–12 → Add relevant internal links, request indexing once, wait.
      ├─ 4–9   → Add the specific missing element (data, experience, examples).
      └─ 0–3   → Commodity content. Rebuild from a different angle, or retire it.

Step 0: Triage Before You Fix Anything

A common mistake is assuming every affected URL should be indexed. Google's Page indexing documentation states that site owners should not expect 100% of their known URLs to be indexed. Generally, only canonical pages with search value need inclusion.

The Index Triage Matrix

Sort every URL in the report into one of four buckets before touching a single page.

A note on PDFs: Google can index PDFs, so do not exclude a file simply because it is a PDF. Exclude a PDF when it is intentionally duplicated by a preferred HTML version, obsolete, private, low-value, or not intended for Search. If both an HTML page and a PDF should exist, clarify their distinct roles and link to the preferred version consistently.

Two rules govern this matrix.

Rule 1: Never conclude a usefulness problem before eliminating technical ones. A July 2026 case reported by practitioners saw an entire site land in this status after a migration. The reported cause was one line in robots.txt, Disallow: /?, intended to block tracking parameters, which also blocked the theme's CSS and JavaScript because those loaded through parameterised URLs. Googlebot was rendering a heading and some boilerplate. Content improvement would not have addressed that.

Rule 2: The Expected bucket is usually the biggest, and it is not necessarily a problem. If your report shows 40,000 URLs and 37,000 are feeds and pagination, your actual problem may be 3,000 URLs. Many people react to the headline number and never reach the real list.

Which pages genuinely need indexing?

Then map each situation to an action:

The metric to track instead of the raw count

Framework #3 (author-created): Valuable Page Index Rate (VPIR). This is an internal prioritisation tool, not a validated Google metric, and its thresholds do not predict Google's indexing decisions. Track it alongside the raw count:VPIR = (indexed canonical pages you intend to rank) ÷ (total canonical pages you intend to rank) × 100The denominator excludes every feed, filter, archive, and utility URL, so you are measuring indexation of pages you actually want in Search.

The bands below are suggested internal prioritisation ranges chosen by the author. They are starting points for triage conversations, not evidence-based cutoffs. Calibrate them against your own site's history before treating any number as meaningful.

Mueller's comments about sitewide quality are the reason for that last row: where a large share of intended-to-rank pages are unindexed, page-level work may underperform. Recompute VPIR on a regular cadence you choose, and adjust the bands using your own data.

The 60-Minute Diagnostic Workflow

Do not guess, and do not work URL by URL. Template-level problems affect every URL using that template, so test three to five representative URLs per template (blog post, product, category, landing page) rather than all of them.

Step 1: Export and filter the report

1. Search Console → Indexing → Pages
2. Under "Why pages aren't indexed," click Crawled – currently not indexed
3. Export (top right) → Download CSV, then open table.csv from the zip
4. In a spreadsheet, add a Bucket column and strip out expected patterns

Patterns worth filtering first:

/feed/          /page/          /tag/           /author/
?replytocom=    ?utm_           ?s=             /wp-json/
?orderby=       ?filter_        ?color=         ?size=
.xml            .json           /search

Do not blanket-filter .pdf; see the PDF note above.

Two limitations of this report:

• It lags. Practitioners tracking it closely report the Page indexing report updating less frequently than the URL Inspection tool. This is an informal community observation, not a documented refresh schedule.
• It samples. Search Console shows and exports a limited set of example URLs, so on large sites you are seeing a subset. For larger URL sets, use the URL Inspection API rather than the report.

Step 1b: Understand the two URL Inspection views

Google's URL Inspection documentation distinguishes two things that people routinely conflate:

• Google Index reports the indexed version of the page as Google currently has it stored, which may be older than your live page.
• Live Test tests whether the currently published page can be accessed and rendered under present conditions.

A successful live test does not prove the URL will be indexed. The live test does not evaluate every indexing, duplication, canonical-selection, or quality signal. When the aggregate report and the per-URL tool disagree about a specific URL, the per-URL tool is the better reference, but read the Google Index panel and the Live Test panel as answering two different questions.

Step 2: Date the problem

Check the last crawl date for each priority URL.

The table below summarises observations from Indexing Insight's customer dataset (1.4 million pages across 18 sites). These are correlations within one commercial dataset, not Google rules, guaranteed cutoffs, or thresholds Google has published. Low crawl frequency and non-indexing plausibly share common causes rather than one causing the other.

Recently crawled pages on healthy sites sometimes index without intervention. Where a URL has been stuck across several crawls, a change in its signals is usually needed.

Step 3: Establish the URL's Search history

Run the 16-month impression check described earlier, remembering that zero impressions do not prove a URL was never indexed. Tag surviving URLs and sort RECOVERY URLs by peak historical impressions. That sorted list is your work queue.

Step 4: The live render audit (Gate 3)

This is the highest-value few minutes in the whole process.

1. URL Inspection → paste URL → Test Live URL
2. Click View Tested Page
3. Open the Screenshot tab
4. Ask one question: can you read your main content in that screenshot?

If the answer is no, stop. You have found a problem to fix before considering anything else.

On soft 404s specifically: Google may classify a page as a soft 404 when it returns a successful HTTP status but appears to be an error page, an empty page, or a page without meaningful primary content. Word count alone is not the deciding factor. Google's technical requirements cover the baseline a page must meet.

Also verify:

• Main content does not require a click, swipe, or login to appear
• The mobile version contains equivalent primary content. Google indexes the mobile version of a page, and its mobile-first indexing documentation states that primary content, metadata, structured data, and image information should be equivalent on mobile and desktop
• Failed API requests do not produce an indexable 200-status error page
• The canonical tag is present in the rendered HTML, not just the source

A page rendering correctly in your browser does not establish that Google rendered it the same way.

Step 5: Robots.txt and HTTP headers (Gates 1 and 2)

Check robots.txt for over-broad rules. Patterns that cause damage in practice:

Disallow: /*?*          ← blocks parameterised URLs, sometimes including theme assets
Disallow: /wp-content/  ← blocks CSS and JS on WordPress
Disallow: /assets/      ← same problem, different stack
Disallow: /*.js$        ← blocks JavaScript on JS-dependent sites

Google's robots.txt documentation makes the division of responsibilities clear:

• robots.txt controls crawling.
• noindex controls whether a crawled page may be indexed.
• If crawling is blocked, Google may never see a page-level noindex, so blocking is not a reliable way to keep a page out of the index.
• Do not use robots.txt to remove an already indexed page from Search.
• Adding noindex does not clean a URL out of Search Console reporting; it usually changes how the URL is categorized.

Separately, a blocked resource can break the render of a page that is itself crawlable. That failure mode can produce large numbers of URLs in this status.

Check for X-Robots-Tag in the HTTP response. A noindex delivered via HTTP header does not appear in your page source and will not show up when you view source in a browser. A CDN can also be configured to serve it selectively.

Check it in URL Inspection → View Tested Page → More Info → HTTP Response, and look for:

X-Robots-Tag: noindex

Verify the status code independently:

bash

curl -I https://example.com/affected-page/

# Follow redirects and see where you actually land:
curl -L -s -o /dev/null -w '%{http_code} %{url_effective}\n' \
  https://example.com/affected-page/

CMS plugins sometimes report a status that differs from what the server sends. Cross-check with an independent status checker.

Check for intermittent server errors:

• Search Console → Settings → Crawl stats, filtered by response code
• Server logs filtered to 5xx responses for verified Googlebot requests
• Any spike in 429 (rate limiting) or 503

Verify Googlebot before trusting your logs. User-agent strings can be spoofed, and a meaningful share of traffic claiming to be Googlebot is not. Google's Googlebot verification documentation describes reverse and forward DNS verification, and publishes IP ranges you can match against. Do this before drawing conclusions from crawl-frequency analysis.

Cross-check with Bing. Submit the URL to Bing Webmaster Tools. If Bing indexes the page and Google does not, basic public accessibility may be less likely to be the only problem. However, the difference does not identify Google's reason for excluding the URL. If neither engine indexes it, a technical cause becomes more plausible.

Step 6: Canonical conflicts (Gate 4)

Duplication can produce this status, because Google does not always assign the dedicated "Duplicate" statuses. Near-duplicates sometimes land here instead.

Check: URL Inspection → Page indexing section → compare User-declared canonical against Google-selected canonical. If they differ, Google has selected a different representative URL. Canonical tags are signals, not directives.

Per Google's canonicalization documentation:

• Redirects are a strong canonicalization signal.
• rel="canonical" is a strong signal.
• Sitemap inclusion is a weak signal.
• Signals can reinforce one another when they agree.
• Google may select a different canonical than the one you declared, particularly when signals conflict.

Google does not publish an exact ranked ordering of every signal, so do not treat any such list as authoritative.

Where Google selects a different canonical, the remedy is often not the canonical tag itself but the surrounding signals, especially internal links. If you declare page A as canonical while your internal links overwhelmingly point to page B, align the links.

Non-obvious near-duplication to check for:

• Location pages differing only by city name
• Product variants (size, colour) on separate URLs
• Category pages surfacing overlapping product sets
• "Best X for Y" articles where Y barely changes the substance
• Programmatic pages built from one template with different data points
• Syndicated or manufacturer-supplied descriptions
• HTTP/HTTPS, www/non-www, trailing-slash variations

Google's guidance is direct that creating many pages primarily for search engines is a spammy practice, and pages competing for the same purpose can behave as duplicates even when the text is not identical.

Step 7: Internal links and click depth

Search Console → Links → Internal links, or crawl your site with a crawler.

• Pages with few or no internal links pointing at them are worth examining first
• Pages buried deep in the site architecture are worth examining alongside them
• Orphan pages that exist only in the XML sitemap are the extreme case

There is no published minimum number of internal links or maximum click depth that Google requires. Relevance, placement, anchor context, and the importance of the linking page matter more than any numeric target. Treat link counts as a comparative diagnostic across your own site, not as a threshold.

Step 8: Zoom out to the pattern

Now stop looking at individual URLs. The pattern across affected URLs is more informative than any single one.

The Differential Diagnosis Table

Framework #4 (author-created): Symptom fingerprints. No single symptom identifies a cause. Combinations narrow the field. Find the row matching your observations to get a working hypothesis, which you then test. These are starting hypotheses, not confirmed diagnoses.

Gate 5 in Depth: The Value Question

If the page returns 200, is indexable, renders correctly, has a clean canonical, and is genuinely unique, page-level usefulness and redundancy become the remaining plausible factors. This is the honest, difficult part of the topic. Note throughout that these remain hypotheses; Search Console does not confirm them for any individual URL.

What Google said in 2026

John Mueller, Search Off the Record, "How to read the Indexing Report" (July 16, 2026), described the sitewide effect: where Google's systems have serious concerns about a website's overall quality, they will reduce the number of pages they index, because spending significant crawling and indexing effort on a site they have strong concerns about makes little sense to those systems. He said this typically results in both less crawling and less indexing.

On how to respond, Mueller advised against treating these situations primarily as technical problems. Where a site shows a broad pattern of pages not being indexed and no technical explanation exists, he suggested stepping back to consider quality across the site rather than debugging individual URLs.

Martin Splitt, in the same episode, widened the definition of quality past the text. He described pages where the text is technically present but is hidden behind ads, interstitials, moving elements, or filler content, and said the full experience of the page has to be taken into account because that is what users encounter.

According to Marie Haynes's attendee notes from Google's 2026 Toronto Search Central Live event, a Google presenter described crawling as downloading the page, with inclusion in a database following only if the content is judged useful. The notes report the presenter giving two reasons a crawled page might be excluded, a technical issue or an assessment that the page was not good, and noting that where large numbers of pages cover an identical topic, Google may determine a given page is unlikely to be useful in Search. Because these are attendee notes rather than an official transcript or recording, treat the specifics as second-hand paraphrase. Google's official announcement of the Canada 2026 event confirms the event itself.

The commodity content problem

A recurring finding in published case reviews is commodity content: material that competently restates what already exists, that many people could produce on the subject.

Haynes's July 2026 analysis of pages stuck in this status puts it directly:

"These pages are usually not junk. They're good, decent articles — as good as the pages that Google is ranking. And that's just the point. The pages aren't special or any more valuable than what currently exists."

This is one experienced practitioner's assessment of the cases she reviews, not a Google statement about your URL. It does help explain why quality checklists can mislead: your page passes them, and so do the pages already indexed.

Framework #5 (author-created): The Index Value Equation. This is a mental model, not a formula with measurable inputs and not anything Google has published. Do not treat it as computable:Marginal index value ≈ (unique information × real search demand) ÷ (redundancy × serving cost)Its only practical use is to focus attention on three levers you control: increase unique information, target demand that is genuinely underserved, or reduce redundancy by consolidating your own overlapping pages. Adding words to a page changes none of them.

The Delta Test: score any stuck page

This is an author-created scoring rubric for prioritisation. The scores have no relationship to any Google system.

Score each question 0, 1, or 2.

Scoring (author's suggested prioritisation bands):

Question 4 deserves its own section

Open the query your page targets. Expand the AI Overview. Read it. Now read your page.

If your page says roughly what the AI Overview says, consider what a searcher gains by clicking through to read the same thing again.

On the measured effects: Ahrefs' click-through-rate analysis found lower clicks to top-ranked results when AI Overviews are present, and an independent randomised field experiment measured a substantial reduction in publisher clicks.

Important boundary: these studies measure click-through rates. They do not show that Google has raised its indexing threshold, and they do not establish that AI Overviews cause more pages to be excluded from the index. Any connection between the two is the author's inference, offered as a hypothesis about why differentiation matters more than it used to, not as a documented mechanism.

What information gain actually looks like

Vague advice to "improve quality" is not actionable. More concrete additions:

• First-hand testing with methodology and results
• Original data: your own survey, your own logs, your own aggregate numbers
• Original screenshots and photography, not stock or vendor images
• A framework or decision tool that does not exist in the current top 10
• Expert commentary attributed to a named, credentialed person
• Case studies with specifics: dates, numbers, what failed
• Templates, calculators, downloadable resources
• Local or regional context that national competitors do not have
• Corrections of a common misconception, with evidence

One genuinely original section is worth more than added filler. Do not invent experience, tests, statistics, or expertise; show how the information was created and cite reliable sources.

Trust signals worth adding

• A named author or reviewer with verifiable credentials
• Clear publication and meaningful update dates
• Citations to primary sources
• A stated editorial or fact-checking policy
• Transparent methodology for any original data
• Accurate contact and business information
• Visible corrections when information changes

The page experience dimension

Following Splitt's point, audit the page as a user experiences it:

• Ad density above and inside the main content
• Intrusive interstitials and consent walls
• Layout shift and slow loading
• Filler paragraphs before the answer
• Main content pushed below promotional sections
• Confusing navigation or misleading headings
• Poor mobile presentation

Google's page experience guidance frames Core Web Vitals and page experience as contributors to overall search success. Passing Core Web Vitals does not guarantee indexing, and failing them does not automatically cause exclusion. Google does not use a single page-experience signal as a universal indexing gate.

The position on AI content

Google does not exclude content from indexing simply because AI helped produce it, and AI-generated content is not automatically disqualified. Google's guidance on generative AI content and its helpful-content guidance focus on usefulness, originality, purpose, accuracy, and compliance with spam policies rather than production method. Scaled content created mainly to manipulate rankings can violate spam policies regardless of whether a human or a machine produced it.

In the July 2026 episode, Mueller made a related point about perception: if most of a website is AI generated and visitors can tell, they may conclude there is nothing unique or valuable available to them there.

Review AI-assisted or programmatic pages for:

• Repeated factual errors or fabricated citations
• Templated paragraphs with swapped variables
• No original contribution over what is already indexed
• Pages generated for every minor keyword variation
• Duplicate examples across pages
• Content unrelated to the site's real audience
• No human editorial review

Framework #6 (author-created): The "Anyone Could Have Written This" prompt. Before publishing, paste your draft into an LLM and ask: "Is this likely to be considered commodity content? What in here could only have been written by someone with direct experience of this subject?" If the answer to the second half is "nothing," you have a signal worth acting on. Follow up with: "Give me 20 ideas that draw on my first-hand experience to make this substantially better than anything else on this topic."This is a pre-publication filter, not a recovery tactic, and it is cheaper than recovery.

The one likely exception

Established sites appear to get more latitude. A page on a domain with strong brand signals and a long track record may be indexed on material that would not be indexed from a newer site. This is a widely reported practitioner observation rather than documented Google policy, and it means benchmarking your content against what large publishers publish can mislead you.

The Fix Playbook, Ranked by Impact

Work these in order. "Request indexing" is deliberately last.

Fix 1: Repair rendering and technical eligibility (Gates 1 to 3)

Cheapest work available, and the most objectively verifiable.

• Unblock the CSS, JavaScript, and API endpoints Google needs to render main content
• Server-render or pre-render primary content rather than relying on client-side hydration
• Remove stray noindex directives, including X-Robots-Tag headers set at the CDN
• Return real 404 or 410 for genuinely gone content instead of pages that read as soft 404s
• Resolve intermittent 5xx, 429 rate limiting, and redirect chains
• Ensure client-side routing returns proper status codes

Fix 2: Consolidate duplicates (Gate 4)

• Redirect true duplicates to the preferred page with a 301
• Add canonicals from variants to the parent product or article
• Remove duplicates from the XML sitemap
• Repoint internal links to the canonical URL
• Align every signal: canonical tags, redirects, sitemaps, internal links, navigation, breadcrumbs, hreflang, and structured data URLs should name the same preferred URL

If two similar pages genuinely both need indexing, they need different purposes and meaningful differences. Changing only the city name, product colour, target keyword, H1, meta title, intro paragraph, or a few synonyms is unlikely to be sufficient. There is no published percentage of difference that qualifies; the test is whether the main content satisfies a different need.

Two location pages that legitimately deserve separate indexing might differ in service availability, local regulations, area-specific examples, original photography, local pricing or delivery terms, genuinely different FAQs, distinct contact details, and verifiable local expertise.

Fix 3: Build internal links from established indexed pages

• Identify your best-performing indexed pages (Search Console → Performance)
• Add contextual links from them to the stuck page, with descriptive anchor text
• Add the page to relevant hub and category pages
• Reduce the number of clicks needed to reach it from your main entry points

Google's crawlable-links documentation explains that Google generally discovers links when they are standard HTML <a> elements with resolvable href attributes:

html

<a href="https://example.com/affected-page/">Descriptive anchor text</a>

Avoid: links that only appear after a user action, links routed through unnecessary redirects, generic anchors used everywhere, links to parameter versions instead of canonical URLs, and mass irrelevant linking intended to manipulate importance.

There is no required number of internal links and no universal click-depth rule. Relevance, placement, anchor context, and the importance of the linking page matter more than hitting a numeric target. Site owners informally report stuck pages indexing after receiving links from established pages, but internal links do not rescue a duplicate, empty, or commodity page.

Fix 4: Rebuild for information gain (Gate 5)

Use the Delta Test to identify what is missing, then add exactly that. Also:

• Sharpen intent match. Answer the target query fully in the first screen, then go deeper
• Reduce boilerplate dominance. Where header, footer, sidebar, and related-posts markup dominates the page, the unique content is proportionally small
• Merge cannibals. Several mediocre pages on adjacent topics sometimes index as one consolidated page, with 301s from the others, when none indexed alone

Fix 5: Prune and consolidate sitewide

Where affected counts run high, run a full content audit.

1. Export all indexed and non-indexed URLs
2. For each thin, overlapping, or zero-traffic page, choose one: improve, merge + 301, noindex (for necessary-but-unsearchable pages), or delete + 410
3. Reduce total indexable URLs until the majority are pages you would want in Search

This feels counterintuitive. It follows from Mueller's point that sitewide quality concerns can reduce how much of a site Google crawls and indexes.

Fix 6: Earn external signals

For newer or lower-authority domains, on-site work alone may not be enough. A few relevant mentions from industry directories, a guest contribution, a supplier or partner link, or genuine PR can support discovery and provide external signals.

Understand the limit: external links can raise a page's apparent importance without changing how Google assesses the content itself.

Fix 7: Sitemap hygiene

Google's sitemap documentation is clear that a sitemap supports discovery and provides a canonicalization hint. It does not guarantee crawling or indexing.

Include: canonical, publicly accessible, 200-status, indexable URLs you want in Search, on the preferred protocol and hostname.

Remove: redirects, 404s and 410s, noindex pages, duplicate parameter URLs, feeds, internal search results, non-canonical variants.

Use <lastmod> only when the page has changed substantially:

xml

<url>
  <loc>https://example.com/affected-page/</loc>
  <lastmod>2026-07-27</lastmod>
</url>

Do not auto-update every date daily. Google uses lastmod when the values are consistently accurate and reflect significant changes.

Fix 8: Now request indexing, once

1. URL Inspection → Test Live URL
2. Confirm the page fetches and the rendered content is correct
3. Click Request Indexing
4. Log the date

Google applies limits to manual indexing requests but does not publish a universal per-day allowance, so treat any specific number you see quoted as unverified. Repeated submission does not guarantee indexing. Do not confuse manual request limits with API quotas: Google's Search Console API limits documentation documents up to 2,000 URL inspection requests per property per day and 600 per property per minute.

Do not misuse the Indexing API. Google's Indexing API documentation limits it to pages containing either JobPosting or BroadcastEvent embedded in a VideoObject. It is not a general indexing tool for ordinary articles, products, categories, or landing pages.

Observation Windows (and How to Tell It's Working)

The ranges below are the author's suggested workflow checkpoints, based on observed cases. They are not Google guarantees, and Google publishes no timeline for indexing decisions. Indexing may occur quickly, take several weeks, or not occur at all if Google continues to exclude the URL. Recheck after a practical observation period based on your site's crawl frequency and size.

Mueller has said that quality reprocessing often takes several months. Haynes's assessment is blunter: for most sites with a large number of pages stuck in this status for quality reasons, recovery will be difficult.

Leading indicators to watch before full resolution

1. Last-crawl dates on stuck pages getting fresher. Google re-checking is an early positive signal.
2. The affected count trending down, and specifically which bucket is shrinking.
3. Newly published pages indexing faster than before.
4. VPIR moving in the right direction, even while raw traffic is flat.
5. Recovered pages earning impressions, confirming indexation translated into eligibility.

Judge trends over periods you set in advance rather than reacting daily.

Crawl Recency: The Early-Warning Signal

Everything above is reactive. You discover a page is unindexed after it is already gone.

Crawl recency lets you observe declining crawl frequency while the page is still indexed and still earning traffic.

Framework #7 (author-created): The 100-Day Alert. Treat days-since-last-crawl as a leading indicator. The 100-day trigger is a threshold the author selected to sit inside the range where Indexing Insight observed indexed rates declining in its dataset. It is not a Google rule and not a validated cutoff. Adjust it using your own site's crawl patterns.

How to build it:

1. Pull lastCrawlTime for your important URLs via the URL Inspection API, within the documented quota
2. Calculate days since last crawl
3. Alert on anything currently indexed but not crawled for an extended period you define
4. Intervene before it changes state: refresh the content meaningfully, add relevant internal links from frequently crawled pages, and confirm it is in a current sitemap with an accurate lastmod

Remember that low crawl frequency and non-indexing plausibly share common causes. Treating crawl recency as a proxy is useful; treating it as the mechanism is not supported.

Fixes by Site Type

WordPress and content sites

Not every WordPress URL reported as unindexed needs correcting. Feeds and thin archives usually have no independent search purpose.

E-commerce

Duplicated manufacturer descriptions are among the most frequently cited catalysts for e-commerce indexing problems in practitioner case reports. Where thousands of product pages differ only by a row in a specs table, differentiation is minimal.

Local and service businesses

Check whether your location pages differ only by city name.

Make them genuinely distinct with: services actually available in that area, local contact details and service boundaries, original project photos and case studies, area-specific regulations or requirements, verified local testimonials, local pricing or response times, and FAQs reflecting real questions from that market.

If you cannot supply meaningful differences, a strong regional hub page may outperform a large set of near-identical city pages.

Programmatic and large-scale sites

• Audit the differences per template. Pages with identical copy and a swapped variable are unlikely to be treated as distinct.
• Index selectively. A smaller set of strong pages may perform better than a large set where most are not indexed.
• Watch for empty combinations. Pages generated from database queries returning nothing can read as soft 404s at scale.
• Note the scaled-content risk. Generating pages to answer every anticipated query variation around a topic carries spam-policy risk. A scaled-content action would not necessarily appear as a manual action in Search Console.

JavaScript applications

Check: server-rendered HTML output, hydration failures, API availability at render time, client-side canonical tags, route status codes, content hidden behind interaction, differences between desktop and mobile output, and blocked scripts or styles. Rendering can be delayed or fail, so important content should be present in the rendered HTML rather than assumed.

Forums and user-generated content

• Delay sitemap inclusion until a thread receives a substantive response
• noindex empty, unanswered, or unmoderated pages
• Remove spam and copied answers
• Consolidate duplicate questions
• Generate descriptive titles from the actual discussion, not the template

New and recently launched sites

Slower indexing on a new domain is commonly reported. There is no domain-age threshold at which behaviour changes, so treat any specific age cutoff you see quoted as unsupported.

• Publish fewer, stronger pages
• Make pages you want indexed easy to reach from your main entry points
• Earn a small number of genuine external links
• Submit an accurate XML sitemap with real lastmod values
• Do not mass-request indexing
• If you bought the domain, check its history

Very large sites

At large scale, some URLs in this status is common. Google indexes a subset of most large sites.

Focus on the composition of the affected set, particularly whether money pages appear in it, and on the trend rather than the absolute number.

Special Situations

"My homepage isn't indexed"

Commonly one of:

• A recently launched domain with few established signals
• A homepage that is a thin JavaScript shell with no crawlable content
• A stray noindex or canonical left over from a "coming soon" or staging setup
• A purchased domain with spam history

Check URL Inspection first, verify the rendered HTML contains your content, and give a new site a reasonable observation period.

"Only my homepage is indexed"

If most important pages are excluded, investigate:

• Sitewide noindex settings (check the "discourage search engines" toggle on WordPress)
• Canonical tags pointing everything to the homepage
• Navigation Google cannot crawl (JavaScript menus, <div> click handlers instead of <a> elements)
• JavaScript rendering failures
• Duplicate or heavily templated content
• Server or CDN restrictions on bot traffic
• A recent migration
• Manual actions or security issues

"My page was indexed and then removed"

Plausible triggers:

• A core or spam update changed how the page is assessed
• Competing content improved
• Your own content thinned through a migration, redesign, or template edit
• An extended gap since the last crawl
• Google selected a different canonical

Compare the current page against the version from when it performed, the previous traffic curve, the Google-selected canonical, and the last crawl date.

"It indexed, ranked briefly, then vanished"

A URL that appears briefly and later drops out may have been reevaluated, canonicalized differently, or affected by changing site and page signals. Search Console alone does not reveal the exact reason. Do not assume it was tested against user engagement data; that is not something Search Console reports or Google has documented for this status.

What NOT to Do: Ten Tactics That Waste Time or Backfire

1. Repeatedly clicking "Request Indexing" on an unchanged page. This queues a recrawl, not a re-evaluation. Google's documentation says there is no need to resubmit. Mueller has noted that needing to submit URLs manually on a regular basis can itself indicate a problem.
2. Padding thin content to hit a word count. No word-count threshold exists. Splitt was explicit that filler content hurts. Adding preamble to a thin page produces a longer thin page.
3. Abusing the Indexing API. It is limited to JobPosting and BroadcastEvent in a VideoObject. Using it for general pages is against documented policy.
4. Buying "instant indexing" services. Anything guaranteeing indexation on a fixed timescale is either misusing the Indexing API or using link spam.
5. Changing the URL slug for a "fresh" evaluation. You lose accumulated signals, create redirect chains, and have not addressed why the content was not indexed. Only consider this after a substantive content rebuild.
6. Deleting and republishing at a new URL. Same problem, plus you reset the page's history.
7. Building backlinks before checking indexability. External links cannot correct an accidental noindex, an empty render, or a canonical pointing elsewhere.
8. Assuming crawl budget explains a small site's problem. Crawl budget is primarily a concern at the scales Google describes in its large-site guidance. It is also not the explanation for this status specifically, since the crawl already happened.
9. Applying noindex to pages you want indexed, to make the report look better. This formalises the exclusion. It also does not remove the URL from reporting.
10. Assuming the Page indexing report is real-time. It updates on a lag and samples on large sites. For decisions that matter, verify per-URL.

Two more worth naming: adding structured data as an indexing fix (valid markup affects rich-result eligibility, not basic indexing), and removing canonical tags to "let Google decide" (canonicals help Google understand duplicate relationships; change one only when it is wrong).

The 90-Day Recovery Plan

The day ranges below are workflow checkpoints selected by the author, not periods after which Google acts.

Days 1–7: Diagnose

• Export the full report; strip expected URL patterns
• Run the 16-month impression check; tag every URL RECOVERY or UNCONFIRMED
• Sort RECOVERY URLs by peak historical impressions, descending
• Live render test on 3–5 URLs per template type
• Audit robots.txt for blocked CSS, JS, and asset paths
• Check HTTP headers for X-Robots-Tag on each affected template
• Compare user-declared vs. Google-selected canonical on a sample
• Review Crawl Stats for 4xx, 5xx, and 429 patterns
• Verify Googlebot in logs before analysing crawl frequency
• Review internal links and click depth for priority pages
• Sort all remaining URLs into the four triage buckets
• Calculate baseline VPIR

Days 8–21: Fix technical and duplicate (Gates 1 to 4)

• Unblock resources needed for rendering
• Remove stray noindex directives, including CDN-level headers
• Address soft-404 signals and thin-template responses
• Resolve server errors and rate limiting
• Consolidate duplicates: 301, canonicalise, or merge
• Repoint internal links to canonical versions
• Add relevant internal links to each priority page from established indexed pages
• Regenerate sitemaps with accurate lastmod; remove non-canonical URLs
• noindex or consolidate bucket 1–2 URLs where appropriate
• Request indexing on genuinely fixed URLs, once each

Days 22–60: Address value (Gate 5)

• Run the Delta Test on your top 20 RECOVERY URLs
• For each 0–3 scorer: rebuild with original material, or retire and redirect
• For each 4–9 scorer: add the specific missing element (data, experience, examples, tooling)
• Merge cannibalising pages; 301 the others
• Add author credentials, primary sources, and honest update dates
• Prune pages with no path to being valuable
• Audit page experience: ad density, interstitials, filler above main content
• Strengthen internal links from your most-crawled pages to priority pages

Days 61–90: Monitor and consolidate

• Track indexed page count and VPIR on a set cadence
• Track days-since-last-crawl on priority URLs against your chosen alert threshold
• Re-check the report: is it shrinking, and which bucket is shrinking?
• Confirm recovered pages are earning impressions, not just indexation
• Document what changed and when, so you can correlate against future algorithm updates
• Set the publishing standard: nothing ships without passing the pre-publish gate below

Tools, Templates, and the Pre-Publish Gate

Free tools that actually matter

The audit spreadsheet

URL | Template | Gate Failed (1–5) | Bucket (Expected/Technical/Duplicate/Value)
| 16mo Impressions | Peak Month | Days Since Last Crawl | Renders OK? (Y/N)
| HTTP Status | Indexing Allowed? | User Canonical | Google Canonical
| Delta Test Score | Internal Links | Click Depth | Recommended Action
| Date Actioned | Status Now | Date Reindexed

Sort by 16-month impressions descending, filter to the Value bucket, and work top-down.

The 60-second pre-publish gate

• Does it contain at least one thing that exists nowhere else?
• Does the AI Overview for the target query already say all of this?
• Is it linked from relevant, frequently crawled pages?
• Is it easy to reach from your main entry points?
• Does the live render test show the full content?
• Is the canonical self-referential and correct?
• Is the main content prominent, not buried under filler or ads?
• Is there a named author with relevant credentials?
• Is it in the sitemap with an accurate lastmod?
• Would I bet money this beats what currently ranks mid-page?

The final pre-request checklist

• Returns a stable HTTP 200
• Is publicly accessible without login
• Is not blocked by a robots meta tag or X-Robots-Tag header
• Has a correct, self-referencing canonical
• Is not a near-duplicate of another page on your site
• Displays complete main content in Google's rendered HTML
• Does not read as a soft 404
• Provides equivalent primary content on mobile
• Serves a clear, single search intent
• Offers something the current top 10 does not
• Cites sources where claims require them
• Has relevant internal links pointing to it
• Is in the correct XML sitemap with an accurate lastmod
• Does not violate Google's spam policies

"Crawled – currently not indexed" reports a state, not a diagnosis. There is no universal trick that clears it, because the same label covers technical faults, canonical selection differences, and judgments about a page's usefulness.

Start by confirming the current status per URL rather than trusting the aggregate report. Decide whether the page should be indexed at all. Then walk the gates in order: access, permission, render, representation, value. A short technical pass either hands you a cheap fix or narrows the field.

For large sites, work patterns rather than individual URLs. Improve the page types that carry real demand, consolidate the duplicates, prune what has no path to being useful, and stop pushing low-value URLs through your sitemaps and internal architecture.

The most productive shift is from asking "how do I make Google index this page?" to asking "what does this page offer that the index does not already have?" That is a harder question, and answering it honestly is most of the work.