Your Rank Tracker May Be Lying: Google’s “Poisoned” SERPs Explained

Some of the strangest SEO news I have come across this year broke this week, and if you pay anyone — or anything — to watch your Google rankings, you need to know about it. Google has been quietly routing

Some of the strangest SEO news I have come across this year broke this week, and if you pay anyone — or anything — to watch your Google rankings, you need to know about it. Google has been quietly routing result clicks through its own redirect layer at google.com/goto — and serving automated tracking tools what researchers call “poisoned” SERPs: normal-looking pages with the results quietly swapped out.

Your rank tracker may be showing you a rankings collapse that never happened. Before you rewrite a page, rebuild your links, or fire your agency, read on.

What is confirmed vs what is reported

Confirmed: Google began testing “goto” redirects in its search results back in July — every result click passes through google.com/goto before reaching the destination page — and has acknowledged them as an anti-abuse measure. That part is on the record. It quietly broke workflows that read destination URLs off the results page, as Teck Hustlers reported this week.

Reported: rank-tracking company Nozzle published deep research on 10 October after analysing 27.5 million SERPs — more than 660,000 keywords across 49 countries, measured between 20 August and 9 October. The timeline: from 13 September, Google requests started failing in large numbers; on 20 September the first poisoned pages appeared; the peak hit on 27–29 September, when 24.9%, 26.7% and 27.0% of collected SERPs respectively were poisoned. Before 13 September the anomaly rate was 0.17%. Afterward it averaged 8.0% — nearly a fiftyfold increase.

What “poisoned” means: an ordinary-looking results page — normal layout, normal HTML, no error, no captcha — with roughly four-fifths of page one quietly replaced by other sources. In normal SERPs, 88.5% of a keyword’s previous sites are still there next time it is measured; in poisoned SERPs, just 20.7%. The replacements are overwhelmingly social and forum sites — Facebook, Reddit, Instagram, LinkedIn and Quora all appear far more often in poisoned pages than in clean ones.

The figure that should worry every business owner: on poisoned pages, the tracked site fell out of the top 10 in 58.6% of cases — against 2.7% in the baseline. And because a poisoned page returns valid HTML instead of an error, the tracker records the “drop” as fact. Your chart shows a collapse that never happened, and that collapse sits in your history forever unless someone corrects it.

Two independent researchers have published matching findings. Maxim Safianov wrote on 14 October that a late-September collapse was probably a fake SERP, and offers a useful three-minute test: compare against a multi-week baseline rather than yesterday, look at the neighbours (if every competitor vanished at once and Reddit and Facebook took their place, it is a substitution), open Search Console, and wait three days. Data backs that up: 80% of poisoned spells lasted a single measurement, and in 77.9% of cases the usual results were back within three days (median two). SEO writer Mediovsky published a detailed breakdown of the same phenomenon.

Opinion — confirmed by nobody at Google: the industry reading, reported by Search Engine Roundtable, is that Google’s real target is AI bots and scraping services hoovering up results at scale, and that SEO tools are collateral damage. Lily Ray put it plainly on 8 October: “rank trackers are too unreliable at the moment.” Glenn Gabe documented the big tools openly disagreeing with each other — one site hit by a spam update was picked up accurately by Ahrefs, late by Semrush, and not at all by Sistrix.

Why this hits Irish local businesses hardest

Local service queries are disproportionately affected — and those are exactly the queries Irish businesses live on. “Emergency dentist Cork”, “roofer Dublin”, “solicitor Limerick”: the searches where a tracker screenshot gets forwarded around the office and a sudden “collapse” triggers panicked phone calls.

A poisoned SERP does not just misreport your position. It makes you act on a problem that does not exist. In my years of local SEO work I have seen all three: pages rewritten that were fine, links rebuilt that were fine, and agencies fired that were fine.

The goto redirect layer plus the poisoning means trackers now need enormous effort to verify a single ranking — Nozzle’s CEO puts it at 500 to 1,000 requests to resolve the links for one ranking. More friction means more tools silently failing and reporting wrong data with a straight face.

A real example from my own audits

Let me give you the pattern I see almost weekly in my managed SEO audits. A Cork home-services business sends me a tracker screenshot: their main “city + trade” keyword has dropped from position 4 to position 28 between Monday and Thursday. Their first instinct is that Google has punished them.

My first move is always the same. I open Google Search Console and look at clicks, impressions and average position for that query over the same window — and compare with the week before. In the case I am thinking of, the tracker’s “collapse” lined up with a week where Search Console showed clicks and impressions completely flat. No cliff edge, no drift. That is the signature of a fake reading, not a ranking drop.

The rule I give every client now: Search Console is the ground truth. Your rank tracker is a weather forecast. You would not cancel a wedding because a forecast three towns over said rain — you look out of your own window first. And this is not the only way Google’s data has been shifting this month — I recently covered Google moving website links higher inside AI Overviews, which is another case of what your tools report drifting from what Google measures.

The checklist I use before I believe a rank tracker

  1. Cross-check in Search Console. A rank tracker “collapse” with flat Search Console clicks and impressions is a tracker problem, not a rankings problem. This is the fastest test there is.
  2. Look at ranges, not days. One-day drops are noise until proven otherwise. Zoom out to two weeks and look for the shape of the change.
  3. Look at the neighbours. If every niche competitor vanished from the top 10 at once and Reddit, Facebook, Quora and LinkedIn took their place, that is the poisoned-SERP signature — a real update never looks like that.
  4. Treat “near me” and local queries with extra suspicion. They are disproportionately affected, so their tracker numbers get the strictest cross-check.
  5. Check whether other keywords fell at the same time. A site-wide “collapse” across unrelated queries is a data glitch; a real drop usually hits a subset of pages.
  6. Ask your tool what it knows. Reputable trackers are publishing about this openly — check status pages and vendor blogs before assuming the data is clean.
  7. Never rewrite pages on rank tracker evidence alone. If Search Console agrees something fell, investigate. If only the tracker moved, wait and recheck in three days.
  8. Move monthly reporting onto Search Console. Build your reports on impressions, clicks and average position from Google’s own data — and keep the rank tracker for competitor watching, where exact positions matter less.
  9. Annotate false dips. If your tracker recorded a collapse that never happened, correct the history or add an annotation. Otherwise you will be explaining that dip on the chart again in six months.

FAQ: rank trackers and Google’s poisoned SERPs

Q1: Is my rank tracker useless now?
No. Trackers still catch trends and remain excellent for competitor comparison. They just cannot be your sole source of truth this month — use them as early-warning instruments and confirm everything in Search Console.

Q2: What exactly are “poisoned” SERPs?
The term comes from Nozzle’s research: Google result pages served to automated scrapers that look entirely normal but carry substituted results — no error, no captcha, nothing that gives them away. Only a keyword’s history reveals the swap.

Q3: Has Google admitted to serving fake results to trackers?
No. Google has acknowledged the goto redirect layer as an anti-abuse measure. The poisoning findings come from Nozzle’s 27.5-million-SERP study and two independent researchers — reported data, not a Google statement.

Q4: Why were local searches hit hardest?
Nozzle’s data shows “near me” and local service queries were affected most. The researchers name no cause, but local SERPs have more moving parts — map packs, local packs, personalisation — which gives automated collection more surface to trip on.

Q5: What should I do this weekend?
Open Search Console. Look at your last 28 days of clicks and impressions for your money queries. If they are steady while your tracker panicked, you have your answer. Then make Search Console the basis of your monthly reporting going forward.

If you want a second pair of eyes on what your real visibility looks like, I offer a free SEO review at kjkseo.com. I will cross-check your tracker against Search Console and tell you, plainly, what is real and what is noise.

share this recipe:

Facebook
Twitter
Pinterest