The Rank Tracker Just Told You It's Missing 80% of What It Used to See

Google has tightened its blocking of scrapers and rank-tracking tools, and one major vendor is reporting an 80% drop in the data it can collect. That's a methodology story, not an AI-visibility story, but it sits directly underneath every AI visibility test that assumes it knows what's already ranking.

What's actually happened

Google made changes around 13 September 2026 that tightened its blocking of scrapers and third-party rank-tracking tools. By 18 September, the effects were visible across the industry. Nozzle reported losing around 80% of the ranking data it was previously able to collect. Sistrix said its data collection was running at a reduced rate. DataForSEO reported similar blocking. Not every tool was hit the same way: reports suggest Ahrefs has continued scraping effectively while Semrush has slowed and Sistrix has fallen behind on detecting ranking changes. One analyst tracking the situation was blunt about it: he didn't have confirmation from every tool yet, and cautioned that data from any of them might currently be wrong.

Why this matters for an AI visibility audit

Set aside for a moment what an AI engine actually answers when someone asks it a buyer question. That test, whether a brand gets named in a ChatGPT, Gemini or Claude answer for a given query, still works exactly as it did before this change. Nothing about how the AI engines generate answers has moved.

What has moved is the layer underneath it. A properly built AI visibility audit doesn't just test AI answers in isolation, it also looks at what already ranks in ordinary Search, because pages that already rank well are more likely to get pulled into an AI-generated answer in the first place. If the tools that report on ordinary ranking positions are currently missing most of what they used to see, that part of the picture gets less reliable, even though the AI-answer test itself is unaffected.

The uncomfortable bit, said plainly

Here's the part that's easy to skip past. If a rank-tracking dashboard is currently missing 80% of what it used to see, and a report built on that dashboard doesn't say so, the report reads exactly as confidently as it did before the data went bad. Nothing about the output looks different. That's precisely the situation where a client, or anyone reading a visibility report, has no way of knowing the ground underneath it has shifted.

The honest response isn't to stop using rank data. It's to say plainly, in any report built partly on it right now, whether the ranking positions quoted have been checked manually or pulled straight from a dashboard that might currently be missing most of what it should be seeing. That's a small addition to a report and it's the difference between a number presented as fact and a number presented as what it currently is.

What this means in practice

For the highest-stakes queries in any audit, the ones a client is going to make a decision based on, verify the ranking position by hand rather than trusting a single dashboard reading. Note in the report which figures were spot-checked and which weren't. And treat this as a live situation rather than a one-off: Google has been fighting scraping for years and the tools generally find a way to adapt, so this is worth revisiting in a few weeks rather than treating as permanent.

None of this changes what AI visibility measurement is for. It's a reminder that the plumbing behind any measurement, AI visibility included, is worth checking occasionally rather than trusted by default.

If you're looking at an AI visibility report right now, ours or anyone else's, it's worth asking exactly this question before you act on it. Our own audits flag which figures were manually checked, precisely because of weeks like this one.