If your rank tracker showed sudden, unexplained drops in late September or October 2026, your rankings may not have changed at all. Google appears to be serving deliberately altered search results to automated systems it suspects of scraping its pages. The SEO industry has started calling these “poisoned” SERPs.
The issue is now confirmed by several major SEO platforms. Ahrefs says it affects between 5% and 20% of search results depending on the data provider, and Semrush has published its own update describing the problem. Google has not commented publicly.
Here is what we know so far, what the data shows, and what SEOs and agencies should do to protect the accuracy of their reporting.
What Are Poisoned SERPs?
A poisoned SERP is a search results page that looks completely normal but contains the wrong results. Instead of blocking a suspected scraper or showing a captcha, Google returns a full page of results that do not match what a real user would see.
This makes poisoning harder to deal with than a simple block. As rank tracking firm Nozzle explained, reported by Search Engine Journal: “A refused request tells a scraper to try again, while a poisoned one gets stored as real.” In other words, bad data quietly enters the database unless a tool is specifically checking for it.
Semrush describes two forms of poisoned SERPs:
- The whole page is clearly off, with results that do not match the query even on page one.
- The top of the page looks normal, but lower results include pages that only loosely match the search.
One example Semrush shared: a Brazilian search for a specific baby stroller model returned a dental supply store’s equipment page, because both use the Portuguese word “carrinho.”
How the Issue Was Discovered
Enterprise rank tracker Nozzle was among the first to publish research on the problem. According to Search Engine Journal, Nozzle’s test flags a first page as poisoned if fewer than half of the sites that normally rank appear anywhere in the collected results. By that measure, 8.7% of first pages were poisoned between September 13 and October 21, 2026, compared with just 0.17% between August 20 and September 9.
Nozzle found that the non-standard SERPs started around September 20 and at one point reached as high as 25% of results. Since October 1, they have made up roughly 5% to 11% of results per day.
What Poisoned Results Look Like
Nozzle’s data, reported by Search Engine Journal, shows clear patterns that separate poisoned pages from normal ones:
| Signal | Normal SERPs | Poisoned SERPs |
| Same sites kept across repeat searches | 88.5% | 20.7% |
| AI Overviews shown | 54.2% | 34.9% |
| “People also search for” shown | 87.2% | 52.7% |
| Image packs shown | 23.0% | 44.2% |
| Forum blocks shown | 20.5% | 27.0% |
Poisoned pages also tend to show more user-generated content from sites like Reddit, Quora, Facebook, Instagram and LinkedIn, and fewer results from sites like Yelp, YouTube and MapQuest. Ads and shopping results appeared fairly stable.
Most poisoning is short-lived. About 80% lasted for a single search, 15% for two consecutive searches, and only 5% for three or more. Nozzle also found that local targeting made no difference: the same keywords tracked nationally and locally were poisoned at nearly identical rates (7.5% and 7.4%).
Why Is Google Doing This?
Google has not explained the change. However, most of the industry believes the target is large-scale scraping for AI systems, not SEO tools specifically.
Nozzle described it as “Google’s response to AI scraping, and rank trackers are caught in it.” Semrush also points to industry discussion suggesting the likely driver is an effort to limit large-scale data collection for AI. Barry Schwartz of Search Engine Roundtable shares the view that Google is mainly targeting AI bots and SERP API services rather than SEO platforms.
According to Search Engine Roundtable, Nozzle’s Derek Perkins described a three-part approach from Google: wide IP blocking aimed at residential proxy networks, more captchas, and injected bad data in suspect SERPs. This follows earlier moves that made scraping harder, including the removal of the num=100 parameter in 2025 and tests of redirect links in search results in July 2026.
As AI search grows, this battle over data is likely to continue. Brands that want to stay visible across both traditional and AI search engines will need to keep an eye on how these platforms collect and display information.
Timeline of Events
- 2025: Google drops the num=100 parameter, making bulk scraping harder.
- July 2026: Google tests redirect links in search results, reportedly to help identify scrapers.
- Mid to late September 2026: Poisoned SERPs begin appearing at scale.
- October 10: Nozzle publishes its research on poisoned SERPs.
- October 12: Search Engine Roundtable reports that third-party rank tracking is not fully working.
- October 14: Search Engine Journal publishes its analysis of Nozzle’s findings.
- October 22: Ahrefs publishes “An update on Google’s poisoning of search results.”
- October 27: Semrush publishes its official update, and Search Engine Land reports that Semrush has confirmed the issue.
How SEO Tools Are Responding
Most major tools have added new checks. According to Search Engine Journal and the companies’ own statements:
- Ahrefs: Uses filters to detect abnormal results, rejects them, requests accurate replacements, and is cleaning up past results that may have been affected. Ahrefs warns that “what works to get accurate data today may not work tomorrow.”
- Semrush: Excludes clearly invalid results during collection, has adjusted its data sources, and added validation checks that flag and recollect suspicious results. Search Engine Roundtable reports that Semrush described the impact as limited and mostly beyond page one.
- SEOmonitor: Estimates about 8% of daily Google results are incomplete or incorrect and re-pulls them before recording. Since October 13, it compares each result against the keyword’s recent history.
- AccuRanker: Discards shortened pages and refetches with a separate browser session. If fewer than half of page one matches recent history, it refetches before recording a change.
- SE Ranking: Corrected ranking data for affected dates and said about 2% of results were still affected as of October 16.
- BrightLocal: Added more data suppliers so one failing provider does not leave gaps, while noting there is “no single fix that solves it completely.”
One important caveat from Search Engine Journal: none of these vendors cite independent testing of their checks, and their percentages measure different things, so they are not directly comparable.
What This Means for SEOs and Agencies
The biggest risk is acting on false data. A keyword that appears to drop out of the top 10 overnight may not have moved at all. Before you rewrite pages or tell a client that their website traffic is dropping, confirm the drop is real.
This matters most for:
- Monthly client reports covering late September through October 2026
- Algorithm update analysis, since some tools caught or missed recent updates at different times
- Competitor research and share-of-voice reports built on tracker data
- Automated alerts and dashboards that pull from SERP APIs
How to Protect Your Ranking Data
- Cross-check with Google Search Console. Search Console data comes directly from Google, so use it to confirm any major drop. Keep in mind that average position varies by device, location and search history. Our guide to fixing Google Search Console issues covers how to read these reports correctly.
- Check your tool’s incident notes. Review how your rank tracker handled the September and October period, including any dashboard notices or status pages.
- Refresh suspicious keywords. Many tools let you re-fetch a keyword. Do this before reacting to a sudden drop.
- Search manually. For high-value keywords, check the live results yourself in a clean browser session.
- Ask about revised data. Ask your vendor whether stored data was corrected, and compare old exports with current data before sending reports to clients.
- Use more than one data source. Combine rank trackers with Search Console, analytics and AI search monitoring so no single source drives decisions.
- Validate API pulls. If you use SERP APIs, ask what safeguards the endpoint applies and whether your own systems check responses.
Conclusion
Google’s poisoning of search results is a reminder that third-party ranking data is an estimate, not a direct feed from Google. Ahrefs, Semrush and other tools are adding filters and cleaning up past data, but no fix is perfect and the situation may keep changing. The safest approach is to treat sudden drops with caution, verify them against Search Console and live results, and avoid making big decisions from a single data source.
Need help making sense of unreliable ranking data or building a reporting process you can trust? Our digital marketing and SEO team can audit your tracking setup and help you separate real changes from data noise.
Frequently Asked Questions
What does “Google poisoning search results” mean?
It means Google is serving altered results pages to automated systems it suspects of scraping. The pages look normal but contain the wrong results, so the bad data can be stored as if it were real.
Does this affect what real users see in Google?
There is no evidence that normal users are affected. The poisoned results appear to target automated requests from scrapers, SERP APIs and rank trackers.
Which SEO tools are affected?
The issue is industry-wide. Ahrefs, Semrush, SE Ranking, SEOmonitor, AccuRanker, BrightLocal, Sistrix, DataForSEO and others have acknowledged it or explained how they handle it.
Has Google confirmed the poisoning?
No. As of this writing, Google has not commented publicly. Most of the industry believes the goal is to limit scraping for AI systems.
Did my rankings actually drop?
Not necessarily. Check Google Search Console, refresh the keyword in your tool, and search manually before assuming a real drop.
