Rank Trackers See Google Ignore Part of the Query

Two rank tracking companies, SE Ranking and Nozzle, report that since mid-September 2026 Google has often returned results that answer only one word of the query, almost always from position 9 down or on page 2. It changes how to read ranking reports for those weeks.

What we know

SE Ranking's study (November 3): it analyzed 10,000 keywords from its research index on Google US desktop. According to the study, 0.09% had an unrelated result on September 16, 57.3% on September 17 and 67.2% on September 24, the highest it saw.
One word: SE Ranking says 84% of the unrelated results were about a single word of the query. Its example: "business bank line of credit" returned the Line messaging app, The Line Hotel and a band, instead of business lending. Single-word queries always came back clean.
Where it starts: per SE Ranking, 96.5% of the unrelated blocks began at position 9 or lower, 3.5% at positions 6 to 8 and none in the top five.
Nozzle's study (October 22): Nozzle says page 2 started dropping query words on September 16 and page 1 on September 20. On desktop, 13.9% of searches it could judge on both pages ignored part of the query, versus 0.06% before, peaking at 39.9% on September 28.
Page 2 versus page 1: according to Nozzle, it hit 12.6% of page 2s and 2.2% of page 1s. On 93.6% of the affected page 2s, page 1 still answered the query.
Short-lived: Nozzle found 85.8% of affected keywords returned to a normal page within three days. The page title still showed the full query 99.8% of the time, with no "Showing results for" notice.
No word from Google: neither study cites a Google explanation, and Search Engine Journal says Google has not explained it publicly.

What changes and what doesn't

It's third-party measurement. Both companies use their own crawlers and different methods, so their percentages are not directly comparable. SE Ranking notes its data cannot show Google's intent, how long the effect lasted within a day or what a person searching by hand saw.
It's not the same as October's "poisoned SERPs." Nozzle separates the two: there, the query was answered but the usual sites were missing; here, the page answers one word. Nozzle says both happen together on 54% of these pages.
Your traffic doesn't change because of it. If the effect is limited to what tools see, your real visits need not have dropped. What may be wrong is the report.
Not all positions are equal. Both studies show the top five almost clean. Keywords ranking from 9 down are the most exposed.

How to tell if you're affected

Check where it fell from. If a keyword suddenly dropped from position 9 or lower between late September and October, it's a candidate. If it dropped from the top five, look for another cause.
Check what replaced you. Open the results your tool saved that day. If they're about one word of your query (another brand, a song, an app with the same name), this glitch is the likelier explanation, not a competitor.
Cross-check in Search Console. In Performance, filter by that query and page and compare impressions and clicks for the same dates, country and device. If they held, you didn't lose real visibility. SE Ranking recommends this same check before treating a drop as real.
Ask whether your tool checks page 2. As Search Engine Journal notes, it helps to know whether your rank tracker reviews page 2 or stops at page 1, since that's where the problem concentrates. In client reports, flag those weeks as unreliable rather than presenting them as a loss.

Sources

Updates: if Google explains the behavior or the tools publish new data, it will be added here with a link.