Ahrefs today told its customers that Google has been "poisoning" search results shown to suspected SERP data providers, and put the share of affected results at between 5% and 20%, depending on the data provider. The Singapore-based SEO software company said it is filtering out abnormal pages, requesting replacements and cleaning up past results that may have been affected.

In Short

Google has been showing different, swapped search results to the automated tools that check website rankings, and Ahrefs now says this hits 5% to 20% of results depending on which data supplier is doing the checking. That matters because agencies, brands and publishers use Ahrefs and similar tools to report on rankings, and some of those reports may include drops that real searchers never saw. Ahrefs says it is throwing out suspicious pages and fixing older stored data, but it has not said how many past results were changed or which dates were affected.

What Ahrefs said

The post, titled "An update on Google's poisoning of search results," appeared on the Ahrefs Product Blog dated October 22, 2026. It is credited to "Ahrefs" rather than a named author, and it runs to six short paragraphs.

"Over the past few weeks, we've witnessed an increase in the "poisoning" of search results by Google, where different results are shown to suspected SERP data providers," the company wrote. "This is an industry-wide issue impacting many companies in our space. Depending on the data provider, we're seeing between 5% and 20% of search results affected in this way."

The response it describes has three parts. "We already have multiple filters in place to detect and reject the abnormal results and request accurate replacements," according to Ahrefs. "We're also cleaning up any past results that may have been affected." Further developments, the company said, will be communicated through dashboard notifications.

The closing paragraph is a hedge rather than a promise. "Search engines regularly make changes, and what works to get accurate data today may not work tomorrow," Ahrefs wrote. "We are continually monitoring the situation, and our aim remains to provide the most accurate and actionable data possible."

That is the entire substance of the statement. There is no methodology, no sample size, no start date, no breakdown by country or device and no description of the filters. The post contains no links to outside research. Google, which has not commented publicly on any vendor's account of the problem, is not quoted.

What is new here

PPC Land has covered the issue twice. On October 13, DemandSphere reported wrong results on about 11% of its daily rank checks. On October 19, Nozzle's study showed Google feeding rank trackers swapped results on 8% of pages, alongside statements from Semrush, SE Ranking, SISTRIX and much lower rates measured by AccuRanker. Ahrefs was absent from those vendor statements. Its post fills that gap and adds four things.

A range, not a rate

The first is the phrase "depending on the data provider." Every other vendor that has published a figure gave a single rate for its own collection. Ahrefs gave a spread, and the wording suggests it is comparing several sources of SERP data rather than one pipeline. The post does not say whether those providers are Ahrefs' own crawling systems, outside suppliers it buys from, or other companies in the market it has observed. Nor does it say which provider sits at 5% and which at 20%.

The range matters because it frames the disagreement between vendors. In PPC Land's own comparison of the published figures, which is a calculation and not a measurement, Nozzle's 8.0% and DemandSphere's roughly 11% fall inside Ahrefs' band, while AccuRanker's rates of under 1% from mid-October fall below it. AccuRanker had argued that the poisoned share tracks how data is collected rather than what is being tracked. Ahrefs' statement that the rate varies by provider points the same way, though it offers no data to test it. The 20% upper bound is also the highest ongoing figure any vendor has stated publicly; Nozzle's peak of 27.0% was a single day, September 29.

Who Google is targeting

The second is the description of the target. Ahrefs says the altered results are "shown to suspected SERP data providers." Barry Schwartz of Search Engine Roundtable and SISTRIX both suggested that Google's real targets are AI companies and SerpApi, with SEO tools caught up in the effort. Ahrefs offers no view on motive. Its wording simply places the companies that sell search data, a category that includes Ahrefs itself, on the receiving end.

Correcting stored history

The third is retroactive cleanup. Earlier vendor statements described checks on new data. SISTRIX said it publishes only results that pass quality checks, and Semrush said it had added "additional checks" to filter results. Ahrefs is the first vendor in PPC Land's coverage to say it is going back over data it has already stored. For customers, that means historical rankings in Ahrefs for the affected weeks may change after the fact. The company did not say how far back the cleanup reaches, how many results it has replaced, or whether it will flag corrected data points in the interface.

Timing

The fourth is the timeline, or the lack of one. "Over the past few weeks" is the only dating in the post. That is vaguer than Nozzle's September 20 first sighting, AccuRanker's September 24 onset and Semrush's "since late summer," and much shorter than AccuRanker's account of smaller rounds of swapped pages going back to January 2026.

Ahrefs' position in the story

Ahrefs had already featured in the practitioner accounts. Glenn Gabe posted on October 7 that "ahrefs accurately picked up the drop" on a site hit during the September spam update, while Semrush was "way late picking it" up. Shoichi Hasegawa, by contrast, wrote on October 1 that Ahrefs numbers did not appear to match reality, a post DemandSphere cited. Ahrefs' statement neither confirms nor disputes either account.

The company has been through this kind of disruption before. After Google dropped the num=100 parameter on September 14, 2025, multiplying the requests needed to collect 100 results, Ahrefs took about six weeks to restore top 100 tracking. That was a cost problem: more requests, more spending, but the data that came back was genuine. The same was true of the goto redirects Google rolled out in August 2026. Poisoning is different, because a swapped page looks complete and normal. A rank tracker that fails to catch it stores a confident, wrong number.

Ahrefs' footprint also extends beyond its own rank tracking. Its SERP data feeds keyword difficulty scores, traffic estimates, competitive analysis and AI Overview tracking products, all of which appear in the footer of the post. The post does not say which of those products draw on the affected data, or whether the cleanup covers them.

Why this matters for marketers

The practical question is which ranking reports from September and October 2026 still hold. Google's spam update ran from September 24 to October 8, producing genuine ranking drops, over the same weeks in which vendors say swapped pages entered their data. A drop recorded in that window could reflect a demotion, a blocked collection or a page Google served only to a suspected bot.

Ahrefs' cleanup adds a further variable. If stored history is being corrected after the fact, a chart exported for a client report in early October may no longer match the same chart viewed today. Ahrefs has not said whether it will log those corrections.

Google's own data does not resolve the problem. Search Console covers only verified sites, and its generative AI report shows impressions without clicks or queries. Competitor positions, share of voice and market-level visibility have always come from third-party web scraping, the activity Google is now degrading. Google has also been fighting that activity in court, with its case against SerpApi still contested after the first complaint was dismissed in July.

Ahrefs has now joined the vendors describing the problem as industry-wide. Its statement that "what works to get accurate data today may not work tomorrow" is a candid one for a company whose product is that data. What it has not done is publish the numbers behind its 5% to 20% range, the step DemandSphere and AccuRanker took. Until more vendors disclose error rates and the dates affected, customers have little way to tell how much of their own data was touched.

Timeline

Summary

Who: Ahrefs, the Singapore-based SEO software company, and Google, which Ahrefs says is showing different results to suspected SERP data providers. Earlier accounts came from Nozzle, DemandSphere, AccuRanker, Semrush, SE Ranking and SISTRIX.

What: Ahrefs said between 5% and 20% of search results are affected, depending on the data provider, and that it is filtering abnormal results, requesting replacements and cleaning up past results. It published no methodology, dates or sample size.

When: The post is dated October 22, 2026, and describes an increase "over the past few weeks." Other vendors date the problem to September 2026 or earlier.

Where: Google Search, as collected by Ahrefs and the data providers it refers to. The post gives no country or device breakdown.

Why: Ahrefs did not give a reason for Google's actions. For marketers, its statement adds a larger vendor to the list confirming the problem, widens the published range of affected results to as much as 20%, and signals that stored ranking history in Ahrefs for recent weeks may be revised.