Enterprise rank-tracking firm Nozzle claims Google has been serving deliberately misleading search results pages (SERPs) to rank trackers and AI scrapers since September 20, with the anomaly reaching 25% of tracked results at its peak and holding at 5–11% per day since October 1. The finding comes from Nozzle's own data — a tool-signal analysis, not a confirmed Google statement — and Google has not commented on the practice.

What does 'SERP poisoning' mean?

Nozzle's term is precise, and the company explains it directly:

"That sequence is why we call it poisoning. Google is not refusing these requests; it answers them with a complete page whose results are wrong, and nothing on the page says so. A refused request tells a scraper to try again, while a poisoned one gets stored as real. It is Google's response to AI scraping and grounding, the bulk collection of its results to feed AI answers, and rank trackers, which collect SERPs the same way, are caught in it."

In other words, the tracker doesn't get an error. It gets a full, plausible-looking page with the wrong results — and stores it as valid data.

Nozzle's theory: Google is doing this to disrupt large language model (LLM) providers that use Google's results to ground AI answers, with rank trackers caught in the crossfire. Search Engine Journal's Roger Montti adds his own speculation that it may also confound spammers reverse-engineering how pages — including AI Overviews — are ranked. That's opinion, not data.

How much ranking volatility does it create?

The numbers are stark. Normal search results retain 88.5% of the same websites across repeated searches for identical keywords. Poisoned results retain only 20.7%. To anyone reading rank-tracking reports, that gap looks like extreme ranking volatility that doesn't actually exist.

The poisoning is also short-lived by design:

• 80% of cases lasted for one SERP crawl of a tracked keyword set
• 15% lasted two successive crawls
• 5% lasted three or more

So a tracker sees a wildly different SERP once, then normal results return — but the bad data is already logged.

Which search features change?

The poisoned SERPs share four traits, per Nozzle: artificial volatility, more user generated content (UGC), fewer AI Overviews, and stable ads and shopping results.

The feature-level data:

• AI Overviews appear in 34.9% of poisoned SERPs versus 54.2% in baseline
• "People also search for" appears 52.7% of the time versus 87.2% in baseline
• Image packs nearly double: 44.2% versus 23.0%
• Forum blocks show more often: 27.0% versus 20.5%

The AI Overview drop has a cost logic behind it — LLM inference to generate an AI Overview costs more than serving ten blue links, so cutting it for suspected scrapers saves money. Nozzle doesn't explain why Related Searches appear less or image packs appear more; the analysis only raises the possibilities that Google is minimizing actionable data or cheapening the page.

Which sites win and lose in poisoned results?

Facebook, Instagram, LinkedIn, Quora and Reddit appear more frequently in poisoned SERPs. MapQuest, Yelp and YouTube appear less often. The analysis offers no explanation for the losers.

The local-search angle matters for anyone tracking GBP (Google Business Profile) and maps visibility. Poisoning hit local searches at roughly the same rate as national ones once variables were equalized. The analysis states it plainly:

"Local targeting makes no difference. City and postal-code SERPs look worse, at 8.4% and 10.7% against 7.8% for whole countries, only because they lean toward mobile and 'near me'. The same phrase tracked nationally and locally, in the same country on the same device, comes out the same: 7.5% and 7.4% across 21,508 such pairs."

That 21,508-pair sample is the study's strongest equalized comparison — local and national queries get poisoned at nearly identical rates.

How widespread is it, and who's affected?

Nozzle found the disruption across countries, industries, and search types, though the rate varies significantly by each. Two side effects also documented: fewer pages of results on some desktop searches, and advertising and shopping units left untouched.

The practical impact splits by audience. Rank-tracking services and in-house SEO teams face diluted, artificially volatile reports. Publishers and brands whose rankings appear to swing wildly may chase phantom changes. And if Montti's spam theory holds, anyone reverse-engineering Google's ranking systems — including AI Overviews — is working from corrupted data too.

One caution on attribution: everything here rests on Nozzle's telemetry from its own tracking infrastructure. Google has neither confirmed nor denied serving differentiated SERPs to automated crawlers, and an alternative read — aggressive anti-bot experimentation — would produce similar symptoms.

Watch whether the 5–11% daily rate holds, creeps upward, or spreads to more request routes, and whether Google comments as more trackers validate the pattern against their own September 20 data.