Nozzle says Google fed rank trackers swapped results on 8% of pages
Rank-tracking vendor Nozzle says Google has been returning swapped-out results pages to automated rank checks. In a study published on 10 October 2026, Nozzle analysed 27.5 million results pages from 49 countries. It found that 8.0% of pages collected since 13 September were "poisoned", against 0.17% before that date. The pages look normal, with no error or CAPTCHA.
On the affected pages, a site that normally ranked was missing from the top 10 in 58.6% of cases, against 2.7% at baseline. Nozzle says daily poisoning peaked at 27.0% on 29 September, and that usual sites returned within three days on 77.9% of affected pages. DemandSphere reported spurious results on about 11% of its daily pages, though it said its organic ranking data was unaffected.
Google has not commented publicly. Nozzle's CEO describes three tactics: wide IP blocking, more CAPTCHAs and injected bad data. Search Engine Roundtable and SISTRIX suggest the target is AI bots and SerpApi rather than SEO tools. Both views are interpretations, not confirmed by Google. Other vendors measure it differently: AccuRanker flagged about 1.47% of its sample from 1 to 9 October.
Why it matters
Many teams judge SEO work by a rank tracker. If a tool stores a poisoned page, a site can appear to drop out of the top 10 overnight and then recover, with no real change in Google.
The timing makes it worse. Google's September spam update ran from 24 September to 8 October, so a ranking dip in that window could have two explanations. Reddit's share of the top 10 also rose from 20.8% to 28.8% on poisoned pages, so competitor comparisons can mislead too.
Vendors are reacting. Semrush says it added checks to filter results and reports limited impact. SISTRIX says it now publishes only results that pass quality checks. SE Ranking said Google had been returning incomplete or unreliable results to trackers. AccuRanker has published its detection method.
What to do this week
- Before you act on a sudden ranking drop, check Search Console impressions and clicks for the same pages. Those are Google's own numbers for real searches.
- Ask your rank-tracking vendor whether it detects and filters poisoned results, and how it reports its error rate.
- Annotate the 13 September to 9 October window in your reports so nobody reads an artefact as a penalty.
- Do not roll back content or links because of one day of tracker data. Wait for a second collection.
- Pair rank data with outcomes: traffic, leads and AI search citations.
Rank trackers are still useful, but treat one sharp move as a question to check, not as a verdict.