Our analysts track the ranking and AI search changes that move agency traffic reports. Brands building search programs can partner with vetted SEO agencies for strategic implementation.
The latest SEO data raises questions about both the reliability of ranking reports and how much search visibility AI systems now occupy.
- Ahrefs said Google showed different results to suspected SERP data providers, affecting 5% to 20% of search results depending on the provider.
- Third-party tracking tools showed ranking volatility on October 20, 21, and 24, and Google has not confirmed an update.
- Semrush found AI Overviews on 97% of the search volume behind 400 commercial keywords in August, and on 88% of the keywords themselves.
- SEO consultant Marie Haynes found 45 of 100 sites skip GA4 key events, which tie organic visits to leads and sales.
- Google analyst Gary Illyes said robots.txt governs crawling, and the standard doesn't dictate what a system does with the content afterward.
Before anyone acts on ranking or AI search numbers this week, the first job is checking where the numbers came from.
Our Take: Which SEO Numbers Skip the Scraper?
SEO has a measurement problem that runs deeper than unreliable rank trackers.
Rankings make for a tidy report, and a week-over-week jump gives everyone something to point to, even when it says little about real performance.
Haynes' analysis shows how deep the problem runs, since nearly half the sites she reviewed had no key events configured in GA4.
These sites may track visibility in detail, and still have no reliable way to connect search activity to leads or sales.
We'd trust the numbers no scraper touches, which means Search Console clicks and the GA4 key events a team defines for itself.
Until trackers settle down, SEO reporting should put a caveat on any ranking change that comes from a questionable source.
Better rank data helps, but a measurement strategy that stops at visibility can't prove SEO ROI.
Ahrefs Sees Up to 20% of SERPs Poisoned
Ahrefs reported on October 22 that Google has been feeding different results to suspected SERP data providers.
These providers collect the results pages that rank reports are built on, which is why the SEO tool maker called it an industry-wide issue.
Rank tracker Nozzle updated its own study on October 21 to include page 2, and now counts 22.6% of the SERPs it checked since September 13 as poisoned.
Nozzle calls a page poisoned when most of a keyword's usual sites have been swapped out.
Most of these swaps reverse within three days, which is why the tracker reads a one-day drop out of the top 10 as likely poisoning.
Ahrefs said it filters out abnormal results, requests accurate replacements, and is cleaning up past data that may have been affected.
Further updates will come through dashboard notifications, which puts the burden on whoever sends the reports.
A rank report already sent to a client may show numbers that Ahrefs has since corrected, and the client never sees the dashboard notice.
These silent corrections get awkward fast when a retainer links fees to ranking targets.
A position won or lost in these weeks may have come from a result that only the tracker saw.
Ranking Volatility Spikes Twice in Five Days
Many third-party tracking tools picked up significant ranking volatility on Saturday, October 24, SEO news editor Barry Schwartz wrote.
Heavy industry chatter about traffic changes came with it.
Google Search ranking volatility is heated 🔥 on Saturday, October 24th https://t.co/qqPvvW02q5
— Barry Schwartz (@rustybrick) October 25, 2026
Together, the two signals make him think Google released a ranking adjustment on Friday and into Saturday.
The spike follows one Schwartz flagged on October 20 and 21, and Google has confirmed neither.
Schwartz also cautioned that the tracking tools behind these readings may still be struggling.
Nozzle counted 35.7% of the SERPs it checked as poisoned on October 21, so some of what trackers logged as volatility may be swapped results.
Trackers also miss real drops, as SEO consultant Glenn Gabe documented on X, with two sites in the same ultra-YMYL niche.
Both dropped out of the results entirely, even for brand queries, and came back on October 16.
Gabe confirmed the drops with the site owners, yet Ahrefs, Semrush, and Sistrix each logged the movement on the wrong dates or missed it.
With no confirmation from Google, a drop in a tracker this week is worth checking in Search Console before it goes into a client report.
AI Overviews Cut Top Commercial Clicks by 70%
AI Overviews appeared on 97.08% of the commercial searches Semrush tracked in August, the company reported on October 21.
Semrush weighted each keyword by its monthly search volume, so roughly 97 of every 100 estimated searches for these terms showed an Overview.
Counted keyword by keyword, Overviews covered 88% of the 400 terms, up from 45.75% in September 2025.
The keywords span banking, software, and eCommerce, each paired with a modifier such as "reviews" or "best X for Y," in Semrush's U.S. desktop data.
The 12% of keywords without an Overview drew less than 3% of the search volume, so the busiest buyer searches are the most covered.
This reach matters because Advanced Web Ranking connects Overviews to more zero-click searches.
In July U.S. desktop data, the top commercial result averaged a 3.51% click-through rate with an Overview, against 11.91% across all results pages.
AWR's own guidance still points SEO teams toward buyer-intent searches as the safer bet because AI Overviews touch them least.
Its figure for commercial queries with an Overview, though, is the lowest of the three intents it charts.
Semrush's numbers suggest that AWR's advice to target buyer-intent searches is aging fast.
Overviews now sit on about 90% of review searches, up from 30% a year ago, so brands win these searches through the sites that Overview cites.
Robots.txt Ends at the Crawl
Google's Gary Illyes described robots.txt as an "access preference protocol" in an October 21 Search Central video.
The episode answered a question about allowing AI crawling while keeping the same content out of model training.
To keep one AI crawler out, he said, a site must name that crawler's user-agent token in the file.
Google's own training opt-out is the Google-Extended token, which keeps content out of Gemini training without affecting Search.
It also refreshed its "control what you share" page on October 21, listing options like the robots meta tag, data-nosnippet, and X-Robots-Tag.
An IETF draft that Illyes co-wrote with Mozilla's Martin Thomson would add a Content-Usage rule to robots.txt.
The rule would let sites state how crawlable content may be used, including for AI training.
But the draft missed its August target for reaching the IETF's steering group and remains a working group document.
For now, robots.txt decides who gets in.
What happens afterward depends on page-level tags for Google Search and on whichever opt-out tokens each AI company chooses to honor.
SEO Industry Insights
Marie Haynes analyzed 12 months of GA4 data from 100 websites with more than 145 million sessions.
ChatGPT sessions equaled just 0.62% of Google organic sessions across that dataset.
On the 55 sites that tracked conversions, ChatGPT visitors converted at 13.8% against 3.02% for Google organic.
Google organic still drove 27 times more conversions than ChatGPT and every other LLM combined.
Haynes also found that ChatGPT referrals stopped growing around the GPT-5 release in August 2025 and have declined since.
These numbers make ChatGPT traffic worth tracking, while Google organic still earns the bulk of any SEO budget.
Chrome developer advocate Barry Pollard checked .jxl files in Google Images on October 23 and found none were actually served as JPEG XL.
Pollard's explanation is that image CDNs often convert files without changing the extension.
Image CDNs often auto convert to a supported format without changing the file extension. It’s confusing, which presumably is why Search docs advise to give the correct file extension where possible.
— Barry Pollard (@tunetheweb.com) October 23, 2026 at 12:34 PM UTC
The check backs Google Search Advocate John Mueller, who said on October 15 that Google Images and Discover previews don't support JPEG XL.
A <picture> element with a standard <img> fallback works for search engines, he added.
Chrome and Firefox now display JPEG XL, but an image SEO plan using the format still needs a JPEG, WebP, or AVIF fallback for Google Images.
Three Checks Before Your Next Rank Report
This week's data problems call for three checks before any rank report goes to a client:
- Recheck sudden drops. Give a keyword that vanishes from a tracker a few days and a Search Console check before reporting it as a loss.
- Set up conversion tracking. Configure GA4 key events to measure whether organic visits generate leads or sales.
- Specify crawler access. Add each AI crawler’s user-agent token to robots.txt if you want to block it from crawling your site.
Ahrefs adds a third vendor rate, but as our last SEO roundup showed, Search Console remains the check that Nozzle and AccuRanker recommended.
Which client report would change if a fifth of the results pages behind it were swapped?
These leading SEO agencies in our directory help brands audit rank reports, set up GA4 key events, and write crawler rules in robots.txt.
