Blog

Poisoned Google Results: Can You Trust Your Rank Tracker?

← All posts
TL;DR
  • Since mid-September 2026, some rank trackers have received Google pages with swapped-out results, so a few reported drops are not real.
  • Vendors say they filter and re-pull bad data. No Google statement had been found as of 30 Oct 2026.
  • Before you change anything, compare Search Console and analytics, then wait to see if the drop repeats.

Mostly yes, with a catch. Since September 2026, some rank trackers have been handed Google result pages that look normal but list different sites from what real searchers see. A sudden drop in your tracker may be bad data, so confirm it before you edit anything. Facts below were checked on 30 Oct 2026.

What happened to Google rank data this autumn?

Rank trackers collect positions by requesting Google result pages automatically. According to Nozzle, a rank tracker that publishes research on this, those requests began failing in large numbers on 13 Sep 2026. From 20 Sep, some pages that did come back had their organic results swapped for other sites. There was no error and no captcha. Nozzle calls them poisoned SERPs.

Nozzle's study (published 10 Oct, updated 21 Oct) counts a page as poisoned when fewer than half of the keyword's usual sites appear. On that rule, 8.7% of first pages were poisoned from 13 Sep to 21 Oct, against 0.17% before. Counting page 1 or page 2, the figure was 22.6% against 0.28%, and it hit 35.7% on 21 Oct after easing in mid-October.

Treat those as one vendor's measurement of its own traffic, not a rate for all of Google. Searches on 30 Oct found no Google statement, and Search Engine Journal's 23 Oct roundup of vendor notices reported none either.

What are the rank tracker vendors saying?

Four vendors have spoken, and they measure different things:

  • Ahrefs (22 Oct): between 5% and 20% of results affected, depending on the data provider. It rejects abnormal results, asks for replacements and is cleaning up past data, without saying which dates (via PPC Land).
  • Semrush (27 Oct): some automated requests return results that differ from what typical users see. It says it has added checks that flag suspect pages for re-collection (via TechWyse).
  • SEOmonitor: recorded incorrect results from 9 to 13 Oct and reprocessed them on 13 Oct (via Search Engine Journal).
  • AccuRanker: re-fetches a page when fewer than half of the previous three days' page-one sites show up (via Search Engine Journal).

The catch: stored history may be corrected after you exported it. Write the date on every export.

What does this mean if you check rankings?

One day of data is weak evidence. Per Nozzle, about four in five poisoning spells last a single result, and 67.9% of poisoned pages had their usual sites back within three days. A real ranking loss tends to stay. Page 2 needs its own look: Nozzle says it is poisoned about 1.5 times as often as page 1, often under a normal-looking first page.

How do you check a sudden drop before changing anything?

  1. Save the evidence. Export the tracker's data with today's date. Vendors are cleaning history, so later numbers may differ.
  2. Compare Search Console. Look at the same page and query over the same dates. Its impressions and clicks come from real searches, so steady numbers there mean the site probably did not vanish. Allow a day or two for data to arrive. For the AI side, see our walkthrough of the Search Console AI Overviews report.
  3. Check analytics. Look at organic visits to the affected pages over the same days.
  4. Read the pattern. Abrupt moves on keywords unrelated to your page (Semrush's warning sign), or forums and social sites crowding the top 10 (Nozzle's), suggest bad data. One page losing one keyword for weeks points at you.
  5. Wait, then act. Give the tracker one or two more updates. If the loss holds in Search Console, investigate the page. If not, send the domain and keywords to your vendor's support. Semrush asks for exactly that.

What should a small team do differently?

Treat tracker positions as estimates. Use Search Console as the record of what happened to your own site, and check a short keyword list by hand each month. Our guide to free AEO and SEO tools covers the free side. Don't swap tools or rewrite pages after one bad week.

What should you watch next?

Nozzle's 35.7% reading on 21 Oct shows the problem had not ended. Watch your vendor's notices, and watch for a Google statement. Until then, a rank tracker is evidence to check, not proof.

FAQ

Does a rank tracker drop mean Google penalized my site?

Not on its own. Check Search Console for the same page and dates first. If impressions and clicks hold, the drop may be bad data.

Did every Google search get poisoned results?

No. Nozzle's figures cover pages it collected for its own rank-tracking customers, not what searchers see.

Can Search Console replace a rank tracker?

For your own site's real impressions and clicks, mostly. It has no competitor data, and its position figure is an average, so it will not match a tracker's daily number.

Sources (checked 30 Oct 2026)