Semrush Confirms Poisoned SERPs Are Corrupting Google Ranking Data

Semrush confirmed on October 27, 2026 that "poisoned SERPs" are contaminating its Google ranking database. The company published a detailed statement explaining that Google now serves some automated search requests results pages that look normal but differ from what a real user sees.

The problem is industry-wide. Every major SEO data provider that collects Google rankings at scale faces the same contamination, according to the statement. Coverage from Search Engine Roundtable points to one likely driver: Google's effort to limit large-scale data collection for AI training systems. Google has made no public comment on the practice.

What does a poisoned SERP look like?

Semrush describes two distinct failure patterns.

• Clearly invalid pages. The entire results page is obviously wrong. Results do not match the query, even in positions on page one.
• Subtly corrupted pages. The page looks normal at the top, but deeper in the results it contains pages that only loosely relate to the search.

The company gave one concrete example. A search in Brazil for a specific baby stroller model returned a dental supply store's equipment page. Both share the Portuguese word "carrinho." That mismatch — a stroller query surfacing dental equipment — illustrates how loosely matched junk results can infiltrate a page that otherwise looks plausible.

This is tool-signal contamination, not a confirmed Google ranking shift. Google has not acknowledged the behavior. The evidence comes from data providers' own collection pipelines and community reporting.

How is Semrush responding?

Semrush outlined three concrete countermeasures in its October 27 statement.

• Its systems detect clearly invalid results during collection and exclude them from the database.
• The company has adjusted its data sources based on data quality.
• New validation checks flag potentially inaccurate results and trigger re-collection of that data.

The company said it continues to refine these processes as search engine behavior evolves. It framed the work as ongoing rather than fixed, an acknowledgment that Google's anti-collection behavior will keep changing.

What should SEOs do right now?

Semrush gave specific advice for users who spot anomalies. If you see unexpected ranking changes — especially for keywords unrelated to your pages — compare against your own first-party data, such as Google Search Console. GSC reflects what Google actually serves, independent of third-party collection pipelines.

The company also asked users to report anomalies. "Contact our support team with the domain and keywords involved, and we'll review your data with you," the statement reads.

For practical purposes, the failure signature is the tell. A real algorithmic shift moves related queries together. Poisoned data produces isolated, nonsensical movements — a stroller page suddenly ranking for dental equipment. Any single keyword showing a dramatic, context-free swing warrants cross-checking before acting on it.

Does this affect AI tracking too?

Semrush positioned the incident within a broader visibility shift. The company said visibility today spans traditional rankings and AI-generated answers, and it will keep investing in both areas.

"We're committed to helping you understand how your brand and your competitors show up across Google's results, AI Overviews, and AI platforms like ChatGPT," the statement says.

That framing matters for the reporting question. If Google is deliberately degrading automated collection, ranking tools and AI-answer trackers face the same upstream problem: the data pipeline itself is under attack, regardless of what the tool measures downstream.

What comes next?

Semrush dated its statement October 27, 2026 and flagged it as provisional, noting it "may be updated as we learn more." The company promised further updates if anything changes for Semrush users.

Three signals deserve monitoring. First, whether Google breaks its silence on serving decoy results to bots — Search Engine Roundtable's coverage attributes the motive to AI data-collection limits, but that remains inference, not confirmation. Second, whether other providers such as Ahrefs, Moz or Advanced Web Ranking disclose similar validation measures, which would confirm the industry-wide scope. Third, whether poisoned results leak into any user-facing Google surfaces, which would turn a tooling problem into a searcher-facing one.

Until Google comments, treat unexplained third-party ranking swings with suspicion and verify against Google Search Console before making decisions.