If your rank tracker showed a sudden, dramatic drop in late September or early October 2026, and your traffic did not move, you were probably not imagining things. Since September, Google has been serving some automated requests a version of its results page that looks perfectly normal but is not what real users see. Rank tracking companies call them poisoned SERPs, and they became one of the most discussed SEO topics of October.
What a poisoned SERP is
A blocked request is easy to detect: the tool gets an error or a captcha, and it tries again. A poisoned result is different. The page comes back complete, with ten results, ads and features, but most of the websites that normally rank for that keyword have been swapped for others. There is no error, so the tracker saves it as real data and your report shows a site that suddenly fell out of the top 10.
What the data shows
The rank tracker Nozzle published the most detailed analysis on October 10, 2026, based on 27.5 million results pages collected for its customers in 49 countries:
- 8.0% of the results it collected from September 13 to October 9 were poisoned, compared with 0.17% in the weeks before.
- The peak was September 29, when 27.0% of the pages collected that day were poisoned. Since October 1 the rate has been around 5% to 11% a day.
- Poisoned pages kept only 20.7% of a keyword’s usual websites, compared with 88.5% normally.
- A tracked site fell out of the top 10 on 58.6% of poisoned pages where it usually ranked, against 2.7% normally.
- It does not last. 77.9% of the keywords checked again within three days had their usual websites back.
- Reddit, Facebook and LinkedIn filled the gap. Reddit appeared in the top 10 of 28.8% of poisoned pages, compared with 20.8% of the same keywords’ other results. Ads and shopping results changed much less.
Other companies saw the same thing. DemandSphere reported spurious results in about 11% of its daily pages on October 7, and SE Ranking, Sistrix, DataForSEO, Advanced Web Ranking and Morningscore all published notices about missing or unusual data. Nozzle also described a wider crackdown: large-scale blocking of residential proxy networks, many more captchas, and then the poisoned results.
An important caveat: Nozzle’s method detects pages where the usual websites vanish and later return. It does not prove what Google intends, and it does not show how often real users saw these pages. Google has not commented on the findings.
Why Google is doing it
Google has never allowed automated scraping of its results, and it has been making it harder for years: it removed the num=100 parameter in 2025, which forced trackers to make ten times more requests for the top 100, and it has tested redirects and new human checks since. The most likely target is not SEO tools but the companies scraping Google to feed AI products, a fight that includes Google’s legal dispute with SerpApi. Rank trackers use the same techniques, so they are caught in the middle. Feeding bad data is also more effective than blocking: a blocked scraper knows it has to try again, a poisoned one does not.
How to track rankings you can trust
- Use Search Console as your source of truth. Google’s own average position comes from real searches, not from scraping, so it is not affected. It is less granular and has a delay of a couple of days, but it is real.
- Never react to a one-day drop. Most poisoned results disappear within three days. Wait for a second and third check before changing anything.
- Cross-check with traffic. A real ranking loss shows up as fewer impressions and clicks. If your tracker shows a crash and your organic traffic is flat, trust the traffic.
- Look at who replaced you. A top 10 suddenly full of Reddit, Facebook and LinkedIn threads, where you normally compete with other websites, is a typical sign of a poisoned result.
- Check the status page of your tool. Most trackers now publish notices when their Google data is affected, and some already filter suspect pages out.
- Track fewer keywords more carefully. A smaller set of commercial keywords, checked several times, is more useful than thousands of daily positions with unknown error.
We compared the main ways to check rankings, free and paid, in our keyword rank checker guide. For finding what to track in the first place, try our free keyword ideas tool.
What it means for SEO reporting
Rank tracking was already losing value: AI Overviews and AI Mode push organic results down, and personalization means there is no single “position 1”. Poisoned results add a new problem, because the error is invisible unless you look for it. For client reports and dashboards, it is time to lead with impressions, clicks and conversions, and to treat third-party positions as an estimate that needs checking, not as a fact.
Frequently asked questions
Frequently asked questions
Are my rankings really dropping or is it my rank tracker?
Compare your tracker with Search Console and with your organic traffic. If the tracker shows a large drop but impressions and clicks are stable, the drop is probably a measurement error caused by blocked or poisoned results.
Which rank trackers are affected?
The problem affects any tool that collects Google results automatically. Nozzle, DemandSphere, SE Ranking, Sistrix, DataForSEO, Advanced Web Ranking and Morningscore have published notices, and users reported unusual data in other major tools.
Is Search Console affected by poisoned results?
No. Search Console reports data from real Google searches, not from scraped pages, so it is the most reliable source for your own rankings.
Do real users see poisoned results?
There is no evidence that they do. The data comes from automated requests, and Nozzle’s study does not measure what ordinary users saw.