Technical

Rank Trackers Face a Data Crisis as Google Poisons Scraped Results

Rank-tracking tools are facing a data crisis as Google returns altered results pages to automated fetchers. AccuRanker disclosed that Google tightened its poisoning detection again on October 15, 20, and 23, poisoning about 40 percent of its first desktop fetches on October 24, while Ahrefs confirmed 5 to 20 percent poisoned results and began rewriting historical data.

How the poisoning works

Google has been sending rank trackers swapped results pages since September, according to reporting on the issue. Research cited in the coverage found that from September 24, Google returned shortened results pages to one automated tracker, affecting up to about 60 percent of its queries at the peak.

The apparent target is unauthorised AI grounding scrapers, which harvest Google results at scale to feed AI answers. But the countermeasure does not cleanly distinguish a grounding scraper from a legitimate SEO rank tracker, so the trackers are absorbing the same treatment.

Why historical data is being rewritten

Ahrefs began rewriting historical data in response to the poisoning, an admission that past tracker readings may also have been affected. That is a deeper problem than a few bad days: if history itself is unreliable, trend lines, before-and-after comparisons, and year-over-year reports all inherit the error.

Agencies should therefore be cautious about restating history from tracker data alone. Where a historical comparison matters for a client decision, rebuild it from Search Console data, which is first-party and unaffected by result-page poisoning.

The sensor problem

Volatility sensors such as Semrush Sensor, Mozcast, and Algoroo measure movement across a fixed keyword sample, and they have reported extreme readings through October, including a 9.5 out of 10 reading for US desktop on October 27. Those sensors are themselves built on scraped results, so extreme readings during a poisoning episode may reflect data corruption as much as genuine ranking movement.

This does not make sensors useless, but it does change how to read them. A sensor spike is a prompt to check Search Console and real traffic, not a diagnosis. During a poisoning episode, the spike is evidence that something is happening to the data pipeline, not proof of an algorithm update.

What changed for everyday SEO work

The practical effect is that rank checks, the cheapest and most common SEO measurement, have become the least trustworthy. Keyword reports that once settled arguments now start them, and SEOs who present tracker positions as precise facts risk misleading clients and themselves.

The adjustment is to move measurement up the chain of evidence: Search Console impressions and clicks first, then analytics conversions, then tracker positions as a secondary signal. It is slower and less granular, but it is grounded in data Google itself reports.

What to do this month

Pause automated rank-change alerts as a trigger for action, and require Search Console confirmation before reacting to any tracker movement. Tell clients plainly that an industry-wide data issue is affecting tracker accuracy, and show them the Search Console comparison so they can see the difference themselves.

Longer term, diversify measurement. Combine Search Console, analytics, AI citation tracking, and link data, so no single degraded source can mislead a decision. The tools that survive this episode will be the ones that treat scraped data as one input among many, not as the ground truth.

A longer view: scraping versus official APIs

The poisoning episode is part of a longer tension between Google and third-party data collection. Google has steadily restricted automated access to its results pages while expanding official channels such as Search Console and its APIs, which give Google control over what data leaves its systems and in what form.

For the SEO tool industry, the strategic response is to build on official data wherever possible and to treat scraped data as a supplement with known error rates. Tools that make that shift openly, showing users where each number came from and how reliable it is, will keep trust. Tools that present scraped estimates as precise facts will lose it.

Sources

Sources: AccuRanker and Ahrefs disclosures, October 22 to 27, 2026, as reported October 28, 2026.

More from Linkbaits

Turn this playbook into links

Linkbaits helps you create content that earns citations and backlinks. Start with the free tools and build your first linkable asset this week.

Start free trial