Rank-tracking tools depend on one assumption: that they see the same search results a real visitor sees. That assumption has been breaking down since around September 13, when Google appears to have tightened its blocks on scrapers, right as practitioners needed clean tracking data to make sense of an already volatile month.
Derek Perkins of Nozzle told Search Engine Roundtable that most of his tool’s Google Search requests are now getting blocked, cutting the volume of retrievable data by roughly 80 percent. He said DataForSEO shows a similar pattern. Sistrix’s status page logged a related issue starting September 16, noting that “further changes” to how Google delivers results had pushed its data collection to a reduced rate. Sistrix’s Steve Paine described the cause as a year-long, currently fast-moving shift in how Google serves results, adding that data is still flowing even as the Toolbox works to catch up.
The timing compounds an already confusing stretch. We covered the practitioner reports of a Sept 4 ranking drop and the renewed chatter that followed days later, and rank trackers are the primary instrument practitioners use to confirm whether a shift like that is real. When the instrument itself becomes unreliable during the exact window it is measuring, teams lose the ability to tell an actual ranking event apart from a scraping failure.
Glenn Gabe’s spot check adds a useful wrinkle: not every tool degraded the same way. Checking one site that fell nearly 100% and then recovered around the Sept 4 update, he found that Ahrefs picked up the recovery, Semrush eventually showed it too, with a lag, and Sistrix did not register the recovery at all. Gabe validated the pattern against the site’s actual traffic, meaning the gap sits in the scraping layer, not in the underlying ranking data.
Search Engine Roundtable’s Barry Schwartz, who compiled the reports, said he suspects Google is currently winning the scraper fight, though he expects vendors to adapt. That reading fits a broader pattern this year: Google has been piloting several ways to verify human traffic, including sign-in prompts and request-verification interstitials, though the goto URL redirects it was using earlier seem to have died down.
The practical stake here is not theoretical. A tracker that silently loses 80 percent of its data rarely throws an error. Instead it shows a flat line, a missing keyword, or a ranking that looks unchanged when it actually moved. During a stretch already marked by disputed core-update signals, that failure mode is worse than an outage because it looks like ordinary, boring data rather than a gap.
Search Console remains the one source in this picture that Google cannot selectively degrade for outside vendors, since it is first-party. Search teams evaluating any ranking movement tied to the Sept 4 through Sept 16 window should cross-check tracker output against Search Console impressions and average position before drawing conclusions, and should treat any single tool’s recovery signal, including Ahrefs, as provisional until a second source confirms it.
Search Engine Roundtable’s Barry Schwartz first reported the scraper-blocking pattern on September 18, 2026, citing Nozzle’s Derek Perkins, Sistrix’s Steve Paine, and Glenn Gabe’s independent data comparison.
