Google Is Breaking Rank Trackers: What Changed in October

If your rank tracker told you your site vanished from Google in the last week of October 2026, your site probably did not move at all. Your tracker broke instead. Starting on October 29–30, Google began throttling the redirect links that rank tracking tools use to read search results, and within a day most results stopped coming back with a real website address attached. One study of 508 search results pages found 58.7% of organic results returning as a Google redirect instead of the site's own URL. A second, separate problem surfaced at the same time: a large share of the results pages served to automated tools had their listings swapped out entirely. Google has said nothing publicly about either change. This article explains what Google actually changed, how bad the damage is according to the published measurements, what still works, and how to check your real rankings while the tools catch up.

What Google actually changed: the goto link

To understand what broke, you need to know how a rank tracker reads Google. It does not ask Google for your ranking. It fetches a normal search results page the way a browser would, then reads the links on that page to work out which website sits at each position. That only works if the links on the page point at the websites they belong to.

Since the summer of 2026, they increasingly do not. Google has been replacing the direct link behind each result with a redirect through its own domain: a google.com/goto address with the destination packed into a scrambled token. The format was first spotted on June 23, 2026 by Alex Greenland, and Search Engine Roundtable described it on July 8. On August 26, a Google spokesperson confirmed the rollout, citing only measures against abuse. Barry Schwartz, the Roundtable's executive editor, read it as anti-scraping aimed at rank tracking tools and AI companies.

For two months, that was mostly an inconvenience. The tools learned to follow the redirect and recover the real destination. DataForSEO wrote on August 27 that only 0.01% of the results pages its API returned could still contain an unresolved goto link — in other words, the problem was solved. The full sequence, from the first sighting in June to the October throttle, shows how quickly that changed. Then, on October 29–30, Google added a throttle. According to DataForSEO's support team, Google now caps how many goto links from a single results page can be resolved in a short window. The first five or six links resolve instantly. After that, each link makes the tool wait: roughly one second for the seventh, three seconds for the eighth, up to 40 seconds by the fifteenth, and in some cases up to 120 seconds per link. The limit applies per results page, not per IP address or query — DataForSEO says changing the IP, changing the query, or waiting before resolving does not help.

The practical effect is simple: a tool that used to read a whole page of ten results in a second now either waits minutes per page or gives up and records "google.com" as the result. Your site is still ranking. The tool just cannot see where.

How bad is it? The numbers from October 30–31

The clearest measurements come from Hasan Kalkan of EverGrow SEO Solution, who published a detailed study of the failure as it happened. On October 30 at 16:34 UTC, a rank tool showed a website as missing from the top 10 for its own brand name. Ninety-seven minutes later, a second query that read the domain Google prints above each result had the same site in first place. The site had not moved. The tool's view of the page had.

Kalkan then measured the scale of it. In a sample of 508 results pages collected across eight countries on October 30, 2,552 of 4,344 organic results — 58.7% — came back with "google.com" as the domain and a goto URL instead of the site's real address. For comparison, the same measurement over October 15–21 had found zero of 71,531 results without a real URL. The change was not gradual. It switched on over two days.

On October 31, Kalkan sent the same queries at the same time to five SERP APIs — the services that fetch Google pages on behalf of rank tracking tools — and compared how many results each one returned without a real URL:

No provider returned every URL. SerpApi opened a public incident on October 30 at 10:39 UTC, announced a fix on the evening of October 31 that restored direct links for the top three results only, shipped a further fix on November 2, reopened the issue on November 3, and shipped another. As of the first days of November, the incident was still open. DataForSEO had posted nothing on its updates page as of November 1, though its support team was replying to customers with the per-page throttle explanation.

The numbers disagree between providers on the exact throttle — Serper resolved only the top three from the early hours of October 30 until the following afternoon, while other tests saw four to eleven links pass — but the direction is the same everywhere: most results below the top three now arrive without a usable URL, and no vendor has a complete fix yet.

The second problem: "poisoned" results pages

The goto throttle is not the only thing making tracker data unreliable right now. A separate study by Nozzle, a rank tracking company, found that Google has been serving some automated tools complete, normal-looking search results pages whose listings have been swapped out. Nozzle calls them "poisoned" SERPs. The pages look genuine — same layout, same formatting — but the results on them are not the ones a real searcher would see. A tool reading one reports that a site dropped, or rose, when nothing changed for real users.

Nozzle collected 27.5 million results pages between September 13 and October 9, 2026, across 49 countries and three device types, and counted 8.0% of the pages with enough history to judge as poisoned. Search Engine Roundtable reported the findings on October 12, and several major vendors — Semrush, SE Ranking, SISTRIX and AccuRanker — have since responded publicly. The full report on Nozzle's poisoned-SERP findings lays out how the swapped pages were detected. Google has made no public statement about the poisoning either.

Put the two problems together and the picture is uncomfortable: since September, roughly one in twelve results pages fetched by a tracker may contain swapped listings, and since late October, more than half the individual results on a page may arrive without a website address at all. A tracker can now be wrong in two different ways at once, and neither failure is visible in the dashboard unless you know to look for it.

What rank tracking tools still work

Not everything is broken, and the situation is not the same for every way of checking a ranking. Here is what still works, in order of reliability.

Google Search Console is the ground truth. It is Google's own record of your site's impressions, clicks and average positions, and it is not affected by any of this — it does not scrape results pages at all. It is free, it covers your whole site, and it is the first place to look whenever a tracker reports something alarming. If Search Console shows your impressions and average position steady while a tracker shows a cliff, trust Search Console. The catch is that it only shows your own site, and its data runs two to three days behind.

Manual checks still work. Open Google in a normal browser, search the query, and look. It is slow and it is personalised, so use a private window, sign out, and treat one manual check as a spot check rather than a measurement — but a manual check cannot be fooled by a goto link, because you are looking at the page the way a searcher sees it.

The top three positions mostly still resolve. Across the October 31 measurements, the best-performing providers were still returning real URLs for the first three results on a page. If your tracker reports a top-three ranking, it is probably right. The damage concentrates from position four downward, which is exactly where most sites live.

Some vendors are adapting faster than others. SerpApi shipped fixes within days, even if partial. Others were still silent on their status pages in early November. Before you judge your rankings, check whether your tool vendor has posted a status update about goto links. A vendor that has acknowledged the problem and shipped a partial fix is giving you better data than one still reporting google.com as a ranking domain.

If you are shopping for a tool right now, the honest answer is that there is no safe pick yet: every major provider was affected on October 31, and the differences between them were a matter of degree. The sensible approach is to keep Search Console as your source of truth, use a tracker for trends rather than single-day positions, and wait to see which vendors recover fully before committing to a new one. A free Semrush alternative stack can cover one-off technical checks without a subscription while the paid trackers sort themselves out — but no free stack replaces a working rank tracker, and nobody should promise you otherwise.

What to do if your rankings "dropped" overnight

A sudden drop in your tracker at the end of October 2026 is more likely to be this throttle than a penalty, an algorithm update, or anything you did. Work through this checklist before changing anything on your site:

1. Check Search Console first. Open the Performance report for the same dates. If impressions and average position are roughly steady, your rankings did not drop — your tracker's view did. Note that Search Console data lags by two to three days, so compare the same date range in both tools.

2. Look at what the tracker actually recorded. If the missing keywords now show "google.com" as the ranking domain, or the URLs are blank, that is the goto problem, not a ranking loss. If the tracker shows a different website in your place, it may be reading a poisoned page — verify it with a manual search before believing it.

3. Check your vendor's status page. Search for "goto" on the vendor's status or changelog page. If they have an open incident, the data is known-bad and will be restated. Do not make decisions on data your own vendor has flagged.

4. Wait a week before reacting. A single bad day in a tracker is noise even in normal times. During an industry-wide data outage it is worthless. No title rewrite, no content change and no disavow will fix a measurement failure, and panicked changes made on bad data create real problems.

5. Check the redirect itself. If you want to see the mechanism with your own eyes, take any Google results page and inspect a result link: the address will start with google.com/goto rather than the site's domain. You can also use a redirect chain checker to see what a URL actually returns — the goto link is, at its core, a redirect, and how 301 and 302 redirects work is the same machinery, just pointed at a different problem.

What this means for rank tracking long term

Google has been tightening the screws on automated access to its results pages for months. The num=100 parameter that let trackers view a hundred results in one look has been removed. Result links were replaced with encrypted goto redirects in the summer. Now the redirects themselves are rationed, and some fraction of the pages served to bots carry swapped listings. Each step is deniable on its own — abuse prevention, infrastructure protection — and Google has confirmed none of them as a policy against trackers. Taken together, they point in one direction: Google is making it steadily more expensive and less reliable to observe its results from the outside.

That does not mean rank tracking is dead. Vendors adapt — they did within days of the throttle appearing, even if only partially — and Google's own data in Search Console is unaffected. But the era of treating a tracker's daily positions as precise truth is over. The tools that survive this will be the ones that are honest about uncertainty: showing which positions were actually resolved, flagging the ones that were not, and leaning on Search Console data where scraping fails. The ones that keep reporting google.com as a ranking domain, or that quietly fill gaps with invented positions, will lose trust faster than any algorithm update could cost them.

For site owners, the practical lesson is the same one Search Console has been teaching all year: own your measurement. A tracker is a convenience, not a source of truth. Your own Search Console property, your own analytics, and an occasional manual check are the three things no throttle can take away. Everything else is a second opinion — useful, but worth exactly as much as its data quality on the day you read it.

Frequently asked questions

Why does my rank tracker show google.com as a ranking website?

Because Google replaced direct result links with google.com/goto redirects, and since October 29–30, 2026 it throttles how many of those redirects a tool can follow per page. When the tool cannot resolve a redirect, it records "google.com" as the domain. Your site is still ranking; the tool cannot see the destination URL.

Did my site actually lose rankings in late October 2026?

Probably not, if the drop appeared suddenly in your tracker at the end of October. Check Google Search Console's Performance report for the same dates: if impressions and average position are steady, the drop is a measurement failure in your tracker, not a real ranking loss. No Google core update was confirmed in that window — the last core update finished in early June 2026.

What is a google.com/goto link?

It is a redirect Google places between a search result and the destination website, with the destination packed into a scrambled token instead of a readable address. Google confirmed the rollout on August 26, 2026, citing measures against abuse. The format itself was not the problem — tools learned to follow it. The problem is the throttle Google added on October 29–30, which limits how many goto links can be resolved per results page.

What are "poisoned" SERPs?

Complete, normal-looking search results pages served to automated tools whose listings have been swapped out, so the tool reports rankings a real searcher would never see. Nozzle, a rank tracking company, counted 8.0% of the pages it collected between September 13 and October 9, 2026 as poisoned. Google has made no public statement about the practice.

What is the best rank tracker right now?

There is no safe single answer while the throttle is active — every major provider was affected in the October 31 measurements, with 55% to 85% of results missing a real URL depending on the provider. The best setup right now is Google Search Console as your source of truth, a tracker used for trends rather than single-day positions, and a vendor that has publicly acknowledged the goto problem and shipped a fix. Revisit the choice once the vendors' fixes are complete.

Are SERP checker tools still useful?

Yes, with the right expectations. SERP checker tools that fetch Google pages face the same goto throttle, so their results below the top three positions are currently unreliable. They are still useful for checking the top of a page, for seeing the layout and features Google shows for a query, and for one-off manual checks. For your own site's positions, Search Console remains the reliable record.