Google Serves Deliberately Wrong Results to Automated SEO Tools

Google has started returning intentionally inaccurate search results to automated scraping requests, corrupting the data that rank tracking tools rely on and leaving SEO professionals unable to trust their own reporting. The change traces to mid September and affects any platform that pulls ranking data by sending bot-driven searches rather than using an official API, which is how most SEO tools operate.

As a senior editor covering this trade day in and day out, it is worth being precise about the mechanics here because the distinction matters to anyone making decisions off this data. SEO platforms do not get ranking information from Google directly. They send large volumes of automated searches that mimic real users and capture what comes back. Google has discouraged this for years and escalated its countermeasures through 2025: removing the parameter that allowed 100 results per request in September, which made tracking roughly ten times more expensive overnight; routing clicks through redirects since July, pushing tools into 500 to 1,000 requests per query; and ramping up "are you human" challenges throughout the year. From 13 September, untrusted requests began failing in volume. From 20 September, the ones that got through started returning different, and wrong, results.

Rank tracker Nozzle analysed 27.5 million result pages across 49 countries in a report published 10 October. Against a baseline of 0.17% abnormal pages, 8% of pages collected after 13 September had most of a keyword's usual top ten results missing, with the correct sites reappearing on a later request. Between 27 and 29 September, a quarter of all pages tracked were affected. On the corrupted pages, organic results retained only 20.7% of the prior page's sites, and AI Overviews appeared on 35% of affected pages versus 54% normally, meaning the tools are also undercounting AI Overview presence. Ads and shopping results held steady throughout.

Australia was hit harder than most markets, with 11.5% of result pages affected compared with 8.5% in the US and 3% in the UK. "Near me" searches, niche queries and local service terms were the most heavily distorted categories, which lines up with where Australian SMB search spend concentrates. Nozzle's worked example involved the query "white label local SEO": a page one dominated by nine relevant vendors on 4 October was replaced three days later, same device and location, by a WordPress development tool, a Jackson Hole steakhouse and a US county health department directory. A browser search for the same term showed the vendors were never gone.

Every returned result is a genuine page, nothing is fabricated. Google appears to be answering untrusted requests with pages that look legitimate to software but would not fool a human reviewer. A flat refusal would simply prompt a retry; a poisoned result instead gets logged as fact and reported to clients as a ranking collapse. Most distortions last a single page and clear within three days, which is part of what makes them difficult to catch in standard monitoring.

Impact

Rank tracking tools are reporting numbers they cannot verify, and clients are seeing apparent ranking crashes that do not reflect reality. The practice adds cost as well as noise: every trustworthy scraped page now comes bundled with several blocked or poisoned ones, pushing up the price of maintaining tracking infrastructure. Google previously closed its official paid data route in January and has pursued litigation against SerpApi, the largest reseller of its results, with copyright claims dismissed in July, refiled narrower, and back in court on 13 October. With the legal and official channels both constrained, degrading scraped data is one of the remaining levers. The data Google cannot corrupt is what it provides directly through Search Console, analytics platforms and actual measured traffic.

What to watch

Professionals should expect rank tracking tools to become both less accurate and more expensive as this pattern continues, reportedly settling at an operating range of 5% to 11% of scraped pages affected through October, well above the pre September baseline. Digital teams and agencies should be able to identify which figures in client reporting are compromised and should be shifting reporting weight toward data supplied directly by Google rather than scraped estimates. The ongoing SerpApi litigation and any further changes to Google's crawling defences are worth tracking, as is whether Google responds to media inquiries on the practice, which it had not done at time of reporting.

Many marketers who rely on an auto backlink tool google may unknowingly receive distorted ranking data, which can lead to poor optimization decisions.

Many marketers who rely on an automated seo backlinks builder google tools may be seeing distorted ranking data without realizing it.

Many marketers who rely on automated rank trackers to see how well they get backlinks google results may be viewing distorted data.

Source: Mumbrella

Related reading from us: 49% of marketers plan to deprioritize SEO, Typeface survey finds