SEO consultant Marie Haynes published an analysis on July 16, 2026 arguing that a growing share of pages parked in the "crawled - currently not indexed" bucket of Google Search Console are stuck there because the content is ordinary, not because it is broken. The pieces she examines are, by her account, competent articles. That, she writes, is precisely the problem.

The status in question sits inside the Page Indexing report in Google Search Console. It marks pages that Googlebot has visited and read but declined to add to the index. It is separate from "discovered - currently not indexed," which flags pages Google knows about but has not yet crawled. The distinction matters because the fixes differ, and because one of the two, on Haynes's reading, rarely has a clean fix at all.

Haynes reports a recent uptick in requests for help with the status. In her words, "I have had a number of site owners reach out to ask for help with indexing issues lately." Across the cases she has reviewed, one pattern recurs. "In almost every case I've examined these pages have quality issues," she writes, describing the material as "commodity content" that rehashes what already exists online "without offering anything new or more helpful than what currently exists."

What Google told attendees in Toronto

Haynes attended the Google Search Central event in Toronto in April 2026. Organizers asked attendees not to attribute quotes to any specific Googler by name, but permitted them to share what was said. Her account of that session forms the spine of her diagnosis.

A presenter explained that when Google crawls a page it downloads the content, and then, in the words Haynes recorded, "If we think it's useful we might put it in a database" - the index. The threshold for what clears that bar, the presenter argued, has moved. Because generative tools have made it trivial for anyone to produce text on any subject, the material Google now wants in its index is content offering two things: personal experience, and knowledge no one else has.

The same session laid out two reasons a crawled page might be left out. One is technical. The other is quality. On the quality path, Haynes quotes the presenter saying a page could be excluded because "we looked at it and found it not to be good," and that when thousands of pages cover an identical topic, Google "might decide that your page is unlikely to be useful in Search." More striking was an admission about how those decisions get tested. The presenter said Google sometimes indexes a page for a period to observe user response: "We are experimenting with seeing which one produces happier users." Haynes calls that "a pretty wild statement."

The rare technical case

Haynes is careful not to collapse every instance into a quality verdict. She recounts a site that had recently migrated and found all of its pages stranded in the status. Using the URL inspection tool in Search Console, she ran a live test and found the rendered page showed only a heading and a few boilerplate words, with the body content missing entirely.

The cause was a single line in the site's robots.txt file: Disallow: /*?*. The rule was meant to block crawling of URLs carrying parameters such as reply or tracking tags. The site's new theme, however, loaded its CSS and JavaScript through exactly those parameterized URLs. The effect was that Google, and every other crawler, was blocked from retrieving most of the page. After the block was removed, Haynes writes, pages began returning to the index slowly.

The episode maps onto a broader point Google has made about crawl controls. PPC Land has documented that Google may index pages blocked by robots.txt even when it cannot read them, because robots.txt governs crawling rather than indexing, and the two are separate systems. The reverse problem, a block that starves the renderer of the resources it needs, produces the empty live test Haynes describes.

She adds a caveat that tempers the alarm. Some pages belong in the status. Feed URLs, paginated URLs, and non-canonical variants are expected to sit there, and their presence is not a signal of trouble.

The quality reading, from Mueller and Splitt

As Haynes was preparing to publish, Google released a Search Off the Record podcast episode titled "How to read the Indexing Report." In it, John Mueller and Martin Splitt walk through the same report, and their framing reinforces the quality thesis.

Mueller drew the line between the two statuses plainly. "Discovered currently not indexed" means "we know they exist, but we haven't actually visited them." "Crawled currently not indexed" means "we visited them and we didn't put them in the index." Asked whether the crawled status signals a quality problem often or only sometimes, he pointed to the systemic case. When Google's systems are "seriously worried about the quality of a website," he said, "they will reduce the number of pages that they index," because sustained crawling of a site the systems have doubts about "doesn't make much sense."

The prescription that follows is not a technical checklist. Mueller framed a large share of unindexed pages with no technical cause as a prompt to reconsider the site as a whole. His argument turns on distance from one's own work: it is "really, really hard to step out of your own perspective," yet the exercise that helps is looking at the site "with the eyes of someone who is not directly involved." Where a site leans heavily on generated text, he said, a visitor may conclude "I can tell this is AI generated" and find "nothing unique or valuable" on offer. He was explicit that this is not a blanket verdict on generated content, but a description of a recurring failure mode.

Splitt widened the definition of quality past the words on the page. Text that is genuinely unique, he noted, can still sit inside "a page that is terrible to access," and he sketched the familiar experience of content "hidden behind ads, hidden behind interstitials," or buried "below a bunch of filler content." Recipe pages carrying a long preamble before the recipe were his example. From Google's vantage, he said, the evaluation has to weigh "the full experience on a page, because that's what users see," since a reader does not "turn on some magic mode that just pulls out the text."

That emphasis on layout and filler is consistent with earlier guidance. PPC Land reported that Google updated its quality rater guidelines with AI content evaluation criteria, adding a section on filler that can "artificially inflate content, creating a page that appears rich but lacks content website visitors find valuable," and a sharper split between a Low rating, where some main content is reused with minimal curation, and a Lowest rating, where nearly all of it is copied with no added value. The same guidelines revision was covered when Google added 11 pages about spam policies to the document, including its first formal definition of generative AI.

Commodity content as the central charge

The term Haynes returns to is commodity content: material almost anyone could produce on a subject, generally restating what already exists elsewhere. Its opposite carries a distinct viewpoint or firsthand knowledge that others cannot easily replicate.

Her framing of the affected pages is where the analysis departs from the usual narrative about indexing gaps. These pages, she writes, "are usually not junk." They are "good, decent articles - as good as the pages that Google is ranking." The trouble is that being as good as what exists is not a reason to add another copy. "The pages aren't special or any more valuable than what currently exists," she writes, which is the whole of the point.

To probe a given URL, Haynes describes a workflow that leans on Google's own systems. She searches queries the page should rank for, expands the AI answer that typically appears, and opens the linked source pages inside a Gemini session in the Chrome sidebar. A saved prompt she calls a non-commodity check then asks Gemini to identify what makes the ranking pages valuable, drawing on criteria pulled from Google's helpful-content documentation: whether content offers original information or analysis, whether it goes beyond the obvious, whether it avoids merely rewriting other sources, and whether it delivers substantial value against competing results. She then turns the same criteria on the client's page. She notes a limit on the method: Gemini has no view into Google's ranking systems and cannot say why a page ranks, so the exercise surfaces qualities that plausibly help rather than confirmed causes. She also flags an exception. Established authorities, she writes, "can get away with a bit more commodity-ness" on the strength of their standing.

The link to firsthand experience is not incidental to Haynes's broader reading of Google. In earlier PPC Land coverage of how Google's AI agent will change search, she traced the reward for demonstrated real-world experience to guidance added in 2022, after which core updates favored sites showing firsthand knowledge, and she noted that rater-guideline changes often precede algorithm changes by months.

Where the June spam update fits

Haynes connects the indexing pattern to a recent algorithmic event, though she is candid about the limits of the connection. She suspects the June 2026 spam update affected a number of sites producing commodity content at scale, and she flags the specific risk attached to the now-common tactic of not merely covering a topic but generating pages to answer every anticipated follow-up query, a practice she warns can invite a scaled-content penalty. Her hedge is direct: "I cannot prove this yet." She also describes the diagnostic difficulty such a penalty creates, since it would surface as an unexplained traffic drop rather than a manual action in Search Console.

That update did land during a crowded stretch of ranking activity. PPC Land reported that Google's June 2026 spam update went live and completed on June 26 after launching on June 24, described in that coverage as broader in felt impact than a typical spam update, and arriving on a cleared runway rather than overlapping a core update as the March cycle had. The policy backdrop tightened earlier in the year, when Google's spam policies were formally extended to cover AI Overviews and AI Mode, addressing scaled content abuse inside AI-generated surfaces for the first time.

The scaled-content concern is not new to the record either. When Google's March 2026 spam update went live, coverage traced the lineage of the scaled content abuse policy back to its March 2024 introduction, part of a run of four broad ranking incidents in roughly thirteen weeks between February and June 2026.

The status was already rising

Haynes's article arrives against a documented, months-long increase in the same Search Console status. In May 2026, PPC Land reported that Google search volatility spiked as pages vanished from the index, noting that the "crawled - currently not indexed" count had been climbing across many properties since roughly early April, following the completion of the March 2026 core update on April 8. That coverage recorded Mueller's prior explanation that the status reflects Google's quality evaluation rather than a technical error, the same position the podcast restates.

A more pointed case reached PPC Land the same month. In an account of a development studio that lost half its search traffic, a twelve-year-old technical blog watched its "crawled - currently not indexed" count grow without explanation and without recourse, its Google traffic down 50 percent in six months. That coverage cited Splitt's August 2024 framing that Google rarely indexes all of a site's content and that server performance and content quality are the two levers, alongside tooling-company data suggesting that 70 to 80 percent of pages carrying the label had been indexed at some earlier point, which recasts the status as active removal rather than a page merely awaiting its turn.

Google's own position on that earlier occasion was covered when the company said that poor indexing on strong hosting indicates quality issues, with the note that a need to submit URLs manually on a regular basis can itself signal a quality problem.

Why this lands on marketers and publishers

The stakes tie back to what indexing is worth in a search environment where a click is no longer the default outcome. A page that never enters the index cannot appear in a ranked result at all, and the value of the ranked result itself has been eroding. PPC Land reported that Ahrefs found AI Overviews cut clicks by 58 percent for position-one content as of December 2025, up from the 34.5 percent the same firm measured in April 2025. A separate randomized field experiment, covered when researchers found AI Overviews cut publisher clicks 39.8 percent, added a causal reading to a body of work that had been correlational.

Set against that backdrop, Haynes's argument sharpens the squeeze on high-volume content operations. She names the exposure directly, writing that she "fears for a lot of SEO agencies" whose primary lever is content creation, precisely because generative tools have made topical coverage cheap. Her reasoning is that if a company can produce content on a topic with generative help alone, "then it's likely not original, insightful and substantially more helpful than what currently exists." She allows for exceptions, citing agencies that use structured pipelines to interview a business, extract its experience, and turn that into original material.

On the difficulty of recovery, Haynes offers little comfort. Where excessive ads or filler are the cause, the remedy is visible. Where the cause is a genuine quality shortfall, the effort required is substantial, and she anchors that word in Google's own documentation, noting that "effort" appears 120 times in the Quality Rater Guidelines. Her suggested route through the problem is to draw on firsthand experience to add to the existing body of knowledge on a topic rather than restating it, and she frames generative tools as a brainstorming aid rather than a substitute for that work.

She closes on a plain read of the direction of travel. "Google does seem to be getting more strict on what it is indexing these days," she writes. For anyone whose traffic depends on pages entering the index in the first place, that is the sentence that matters.

Timeline

  • February 2017: An undocumented Google algorithm change, later cited by Haynes, begins favoring firsthand-experience signals on health and finance content.
  • 2022: Google's Helpful Content system is introduced, using user signals to train ranking on what counts as helpful; guidance on demonstrating real-world experience is added the same year.
  • August 2024: Martin Splitt explains the discovered and crawled not-indexed statuses, citing server performance and content quality as the two primary factors.
  • 2024: The Helpful Content system is retired and its learnings folded into core ranking.
  • April 2025: Ahrefs measures a 34.5 percent click reduction for position-one pages when AI Overviews appear.
  • April 8, 2026: The March 2026 core update completes.
  • April 2026: Haynes attends the Google Search Central event in Toronto, where presenters describe the indexing threshold and the experiment of temporarily indexing pages to gauge user response.
  • Early April 2026 onward: The "crawled - currently not indexed" count rises across many properties.
  • June 24, 2026: The June 2026 spam update launches.
  • June 26, 2026: The June 2026 spam update completes.
  • July 16, 2026: Google publishes the Search Off the Record episode "How to read the Indexing Report," and Haynes publishes her analysis the same day.

Summary

Who: SEO consultant Marie Haynes, drawing on statements from Google representatives at a Search Central event and from John Mueller and Martin Splitt in a Search Off the Record podcast.

What: An analysis arguing that most pages stuck in the "crawled - currently not indexed" status in Google Search Console are held out for quality reasons rather than technical faults, with commodity content, material that competently restates what already exists, identified as the central cause. Haynes also connects the pattern, without claiming proof, to the June 2026 spam update and the risk facing scaled content operations.

When: Published July 16, 2026, the same day Google released the podcast episode on reading the indexing report; the Toronto event Haynes cites took place in April 2026.

Where: The diagnosis centers on the Page Indexing report in Google Search Console and applies globally to any site whose pages are declined for the index.

Why: Pages that never enter the index cannot rank, and the value of ranking itself has fallen as AI Overviews absorb clicks that once reached websites, which makes an indexing threshold that Haynes describes as tightening a direct concern for publishers and marketers who depend on organic visibility.