Rank-tracking tools have detailed how they defend against what they call poisoned Google search results, amid signs that Google serves different results to suspected automated scrapers. According to Search Engine Journal, which published the vendors’ statements on October 23, Ahrefs, AccuRanker, SEOmonitor, BrightLocal and SerpApi each described how they detect suspect result pages, re-fetch them and reprocess historical data. SEOmonitor estimates that around 8% of the results Google returns each day are incomplete or incorrect and get re-pulled before recording, while Nozzle measured 8.7% of first pages as poisoned between September 13 and October 21. Google has made no statement as of October 22. The disclosures matter because late-September and October ranking data may be unreliable for client reports.
What the vendors check before storing a result
AccuRanker eliminates shortened Google results pages, then re-pulls with a fresh browser session. It compares the new page-one results against the sites listed there for that keyword on at least two of the previous three days. If fewer than half appear, AccuRanker re-fetches; if the second pull shows the same new sites, it records a genuine change. About one in 20 desktop pages and one in 12 mobile pages lacks enough history to run the check.
Around 8% of Google’s daily results are incomplete or incorrect and get re-pulled before recording, SEOmonitor estimates. On October 13, it began checking each result against the keyword’s own recent history.
Ahrefs says filters detect and reject abnormal results and request replacements. It estimates that 5% to 20% of search results have been affected over the past few weeks, depending on the data provider, without saying whether that share is measured before or after its filters run. SerpApi, which sells Google results through an API, deployed an October 14 update it said “should reduce the number of instances in which poisoned results are returned.”
What customers see while a result is unconfirmed
What a user sees depends on the vendor’s validation and display rules. For some unusually large same-device drops, such as a fall from rank 4 to rank 14, SEOmonitor “may keep showing the previously recorded rank for up to two daily updates” until it determines whether the lower rank persists.
AccuRanker says that when its history check applies and a page has lost most of its normal sites, a secondary fetch runs before anything reaches the customer’s dashboard, and users can manually refresh a keyword for a fresh fetch. The AccuSERP API returns the page Google sends for a single search, so it skips the history check; it draws first on sessions that passed the rank tracker’s check.
How the vendors revised results already recorded
SEOmonitor says it reprocessed all keywords affected by incorrect results for September 25 through 28. It also says it blocked erroneous data from entering its product until October 8; from around October 9, many bad results no longer carried the signals its quality check depended on. Around 10% of monitored keywords saw incorrect results between October 9 and 13, though the company says most movements in its Visibility metric were minimal.
AccuRanker runs a “sandwich check” over stored keyword data, flagging days when fewer than half the page-one sites match the day before and the day after. On a fixed one-in-200 keyword sample, the flag rate reached 30.2% of desktop keyword-days on September 25 and fell to 0.58% on October 17. The blog post does not say whether flagged days are revised inside customers’ stored data.
BrightLocal CEO Myles Anderson wrote on October 16 that the company is adding data suppliers, so another is available if one fails. He said there is “no single fix that solves it completely” for BrightLocal or anyone else in the industry.
Why It Matters
The poisoning issue changes how agencies should read their own data before acting on it. Teams reviewing potential late-September or October ranking drops should first check how their vendor reported out during those weeks, because a dashboard drop may be a data-quality artifact rather than a ranking change — adding to the search ranking volatility SEOs already face. SEOmonitor’s safeguard can show the previous rank for up to two days of updates while it verifies a big drop. AccuRanker lets users refresh a keyword manually for a fresh fetch.
Search Console remains useful as independent evidence in a dispute about a drop. Google’s own performance reporting documentation notes that average position is based on the site’s highest-ranking result, and that results vary by device, location, time and search history. The episode also shows how much of Google’s surface is shifting: the company has been moving links higher inside AI Overviews and AI Mode while fighting scrapers on the classic result page. Separately, two German courts have begun holding Google liable for false claims in its AI answers, adding legal pressure to the accuracy debate.
Background
Rank trackers have long suspected Google serves different results to automated collection systems than to real users, a practice the industry calls result poisoning. The latest wave became visible in September 2026, when several vendors reported anomalous data. Search Engine Journal’s October 23 roundup is the first side-by-side account of each vendor’s handling of suspect results before, during and after recording them. None of the vendors cites independent testing of how well its checks work, and each vendor’s figures measure different things, so the numbers are not directly comparable.
Source: Search Engine Journal
0 Comments