Google has been handing rank trackers complete-looking results pages with the organic listings swapped out. Nozzle published the data on October 10; Ryan Doser recomputed it from Nozzle's five public CSVs on October 14. Poisoned window (Sept 13 to Oct 9) compared with baseline (Aug 20 to Sept 9): Share of judgeable SERPs: 8.0% vs 0.17% Tracked site missing from top 10: 58.6% vs 2.7% Previous organic sites retained: 20.7% vs 88.5% Ads retained: 83.7% vs 70.6% AI Overview present: 34.9% vs 54.2% Reddit in top 10: 28.8% vs 20.8% Daily rate swung from 0.04% to 27.0%, peaking September 29. Canada averaged 10.5% and had the worst single day at 48.1%. The UK ran lowest at 3.0% and stayed below the US every day from September 20 to October 9. Desktop, iPhone and Android were all hit about equally, 7.1% to 8.3%. US legal was worst at 15.9%, pets least at 4.3%. Recovery is fast, which is why it's dangerous. 80% of spells lasted one collection, and 77.9% of poisoned pages had their usual sites back within three days. Some clients get these automated ranking reports every month, nobody looks at an actual SERP, and the tool records a cliff and a bounce that nobody searching ever saw. It also sits inside the September 2026 spam update, September 24 to October 8. The rise started September 20, four days before the update. A fake drop and a real penalty look the same on the chart. Read the evidence honestly before you quote it to a client Nozzle is a rank tracker publishing research about rank trackers. Doser's page reproduces Nozzle's arithmetic from the same files, so it is not independent confirmation. The poisoned definition depends on the usual sites coming back, so it cannot prove Google did it deliberately. DemandSphere separately estimated about 11% on October 7 using a different measure, and AccuRanker's own validation puts October 1 to 9 at just 1.47%. Google has said nothing. Treat the exact percentage as unsettled and the direction as real. The protocol I'd run on every account right now

Pick a service or product page you want more people to visit. Start at your homepage and try to reach it through the navigation. Where does the route get confusing? An unclear menu label, a missing link, or an unexpected destination? Share the page and one change you’d make.

Our friend Craig Campbell has shared a walkthrough of Backrefs, a new automated link-building tool. In the video, he shows how to set up campaigns, use website content to populate campaign details, and review the links created. He also covers local SEO campaign options and connections to AI tools through MCP. If link building is part of your work, take a look at what he’s putting together: https://www.youtube.com/watch?v=iv_LKaqdHZc What would you want to check before testing a tool like this: the sites it uses, content quality, control over placements, or how you measure the outcome?

A conference talk, webinar, or article can leave you with more ideas than you have time to try. Before adding another task to the list, ask: • What problem would this solve? • Does it fit the website and audience I’m working with? • What would I measure? • What would make me stop the test? Share one SEO idea you’re considering and why it deserves a test. If you’ve ruled one out, what made you pass?

Think of one SEO change you tested recently. It could be an internal link, a rewritten page, a Google Business Profile update, or a technical fix. Share four things: • What problem were you trying to solve? • What did you change? • What happened, and over what period? • What would you do differently next time? A result that stayed flat is useful too. Include the evidence you checked, and mention other changes that might have affected the outcome. No polished case study needed. One specific test and an honest lesson are enough.