Google is poisoning 8% of rank tracker SERPs, Nozzle finds
Derek Perkins says Google now serves normal-looking results pages with swapped rankings to suspected scrapers.
Nozzle's Derek Perkins published a deep dive this morning showing that 8.0% of SERPs returned to his rank tracker come back poisoned. Barry Schwartz reported the finding at Search Engine Roundtable on 12 October 2026.
Perkins describes what Google has been doing since September 2026 as serving complete, normal-looking pages whose results have been swapped out. Nothing on the page shows the swap. Tracked sites fall out of the top 10. Social sites and forums fill the gap. The poisoning passes within days.
Perkins says Google is using three methods against scrapers. The first is wide IP blocking aimed at major residential proxy networks. The second is a sharp rise in captchas. The third is injecting bad data into SERPs that Google suspects are being scraped.
Why the third method is the dangerous one
Blocked requests are easy for a tool to notice and flag. A poisoned SERP arrives looking fine. A tool with no way to detect the swap will keep reporting wrong numbers to customers.
Schwartz lists tools that have posted notices about recent problems, including Sistrix, DataForSEO, SE Ranking, Advanced Web Rankings and Morningscore. Most of those notices address the goto redirects Google began testing in July, which Perkins says are the easier part to handle.
Practitioners quoted in the piece report the mismatch directly. Glenn Gabe documented one site that dropped nearly 100% and recovered on 4 September: Ahrefs caught the recovery, Semrush lagged, Sistrix did not show it. Lily Ray wrote on 8 October that rank trackers are too unreliable right now to share findings from. Alex Wolk said spot-checking in an incognito window is the only way he trusts a movement.
What to do before you report a ranking change
We think Schwartz is right that the injection method matters most, because a tool reporting confident wrong numbers is worse than a tool reporting an outage. The practical check is already in the thread Schwartz quotes.
Before reporting a drop or a recovery to a client this month, cross-check the tracker against Search Console impressions for the same query. Then run a manual check in a clean browser. Daniel Skov Jacobsen's advice in the thread is to look at the GSC trend first and treat daily swings in trackers as measurement noise until confirmed.
Schwartz notes Google's own view is that these measures target AI bots and SerpApi rather than SEO tools. The effect on rank trackers is the same either way. Search Console remains the one first-party source, though it does not cover clicks or queries for AI Overviews or AI Mode.