Poisoned Google SERPs hit 8% of captures—here's a DIY verification pipeline for small-scale SEO

VeryWellVersed · x · 2026-10-19

Building on Derek Perkins's / iPullRank's research on poisoned Google SERPs (Google serving scrapers convincing pages with wrong results), the author shares an approach that works for sub-million SERP scales.

Key facts from the accompanying infographic:

Scale: 8% of 17M SERPs judged poisoned (Nozzle, Sept–Oct 2026), peaking at 27–34% in late September; 80% of poisonings last a single fetch and most keywords recover within 3 days. The real damage is rank trackers charting and alerting on fake drops.
Detection: flag pages with same template/ads but different results; verify via (1) fresh browser identity per fetch with captcha solving, (2) scoring six symptoms (off-topic top-10, unfamiliar domains, shallow pagination, fewer results, no AI Overview, etc.) with <0.5 = suspect, (3) comparing against a baseline built only from domains clean in 60%+ of captures, (4) re-fetching with a fresh session cookie.
Ops: up to 5 independent captures per query; an hourly canary set of head terms tracks baseline, blocks captchas, and retains hourly HTML of flagged pages.

The author notes reliable verification now seems impossible without storage and multiple scrapes, and invites other solutions.

Original post →

Poisoned Google SERPs hit 8% of captures—here's a DIY verification pipeline for small-scale SEO

More from coding & agent