Google's GoTo URL Change: What SEOs Need to Know
Google has started routing clicks on search results through a google.com/goto redirect, which is a Search infrastructure change rather than a ranking update: it degrades what rank trackers, SERP scrapers, and search-data APIs can collect, so treat sudden gaps or jumps in your reports as a measurement problem until you have confirmed otherwise.
Google has changed what happens when someone clicks a result. The link no longer points straight at the destination — it points at Google, which forwards the visitor on. Result clicks now pass through a google.com/goto hop before they reach the site.
Nothing about your rankings changed. What changed is who can still read them. Anyone reading position data from a third-party tool — a rank tracker, a SERP scraper, a reporting dashboard — now depends on software that has to follow that extra hop before it knows where a result leads. The businesses being ranked are not the affected party. If your monthly report comes out of one of those tools, the Google GoTo URL change is your problem.
What exactly changed in Google’s search results?
Until recently, an organic result on a Google search page carried the destination URL in its href. Click it and the browser went to the site. Google now inserts a google.com/goto redirect in between: the click lands on a Google-controlled URL first, and Google forwards the browser on to the destination.
Before
google.com/search?q=... ──click──► example.com/page
After
google.com/search?q=... ──click──► google.com/goto?... ──► example.com/page
└─ Google-controlled hop
a third-party tool has to follow
before it knows the destinationThe extra hop is close to invisible to a person, because the destination page still loads. It is not invisible to software reading the results page from the outside, which finds a Google URL where the destination URL used to be and has to resolve it before recording anything. The change is one added link in the chain, at exactly the point where measurement happens.
Google confirmed the change publicly in late August. Search Engine Land reported that confirmation on or around 26 August 2026. Search Engine Roundtable carried the same confirmation and had been documenting the Google GoTo URL passthrough in testing from July 2026 — that earlier date is Roundtable’s own, and Google did not name it.
Why is Google doing this?
Google’s own explanation, as quoted in both reports, is about abuse rather than about data. The company says it has a long record of deploying technical measures against abuse in Search, that the forms that abuse takes keep changing, and that measures like this one protect its services and its users. That is the whole of the stated reason; Google has published no technical rationale for the GoTo URL update beyond it.
The reading circulating in the industry is different, and it is worth labelling as what it is: an inference by the reporters covering the change, not a statement from Google. That reading is about data access. A Google-controlled hop makes automated SERP scraping harder and more expensive, because a scraper that only reads the results page no longer gets the destination URL for free. Google has said nothing to confirm that motive, so treat it as plausible rather than verified.
What did I see when I tested it?
I checked this myself, in the plainest way available: I ran the same queries twice in the same browser, once signed in to my Google account and once in an Incognito window.
Signed in, the result links carried the passthrough. Hovering a result showed a google.com/goto URL in the status bar rather than the destination, and copying the link address gave me the same thing. In Incognito, on the same queries, I did not see it consistently — some results still handed me the destination URL directly. I did not find a pattern that explained which ones, and I did not try to force one by clearing cookies or switching networks.
If that holds generally, it means the redirect is not applied uniformly, and a tool’s exposure to it may depend on how it fetches the page. I would not build a conclusion on that. I ran those checks in the week of 1 September 2026: one person, one machine, a handful of queries — a manual observation, not sampled or measured data.
What does this break for SEO tools?
Rank tracker accuracy depends on one assumption: that the tool can read a results page and match each position to a destination domain. The redirect removes that certainty. A tool that cannot resolve the hop cannot say whose page sits at position four, so it records nothing, guesses, or drops the row — and none of those outcomes looks like an error in a client report. They look like a ranking change.
The surfaces exposed are the ones that read Google from the outside:
- Rank trackers. Daily position histories built by parsing fetched results pages.
- SERP scrapers. The collection layer underneath most of the above.
- Data providers. Vendors reselling parsed SERP data, so one collection problem propagates widely.
- Reporting systems. Dashboards that read from any of the above and rarely flag missing data as missing.
- Search-data APIs. Endpoints returning positions and URLs, where a null is easy to mistake for a zero.
Ahrefs posted a notice in its own interface while this was unfolding:
Google keyword ranking data disruption. Because of recent changes introduced by Google, you may notice temporary changes in ranking data. We’re working on a solution and will keep you updated.
I saw that notice in the Ahrefs interface in the week of 1 September 2026, and I am reporting it from my own reading of the message: I did not capture a screenshot, so take it as a note of what the tool said rather than as published evidence.
Is this a ranking change or an infrastructure change?
Infrastructure. Nothing here touches the ranking algorithm. Which pages Google considers relevant for a query, and the order it puts them in, are decided before any of this matters; the redirect changes what happens after a person clicks, which is plumbing rather than judgement.
That distinction is the whole practical point, because the failure mode is a category error. A gap in a rank history is a measurement artefact, and a measurement artefact read as a ranking movement produces exactly the wrong response: a rewrite of pages that were fine, a panic call about a competitor who did not overtake anyone, a strategy change made on evidence that was never collected. The data got worse. The site did not.
What should you tell clients about their reports?
Tell them before they ask. A client who reads a report with holes in it and hears the explanation two weeks later has already decided what the holes mean, and no amount of accurate context lands after that. A short note naming the redirect, saying it affects data collection rather than positions, and pointing at the numbers that still hold is a five-minute job that saves a difficult meeting.
Then be careful about what the report leads with. If a rank grid is the only evidence a client has ever been shown, a fortnight of missing rows reads as work that stopped — which is why a conversation about what local SEO actually costs goes better next to lead numbers than next to position screenshots.
For the local SEO work we do, the monthly report leads with calls, forms, and booked jobs, and position data is context underneath them rather than the headline. That ordering costs nothing when the tools are healthy and it costs nothing now, which is a fair test of whether a metric belonged at the top in the first place.
What should you measure while the data is unreliable?
Use the sources that never depended on reading a results page from outside. Google Search Console is reported by Google about your own property, so impressions, clicks, and query-level data arrive regardless of what a scraper can resolve; its position figure is an average rather than a daily rank, which is a limitation worth stating rather than a reason to ignore it.
Analytics carries the rest of the picture: organic sessions, which landing pages receive them, and what those visitors do next. Profile views, calls, and direction requests come from Google’s own reporting on the listing, and the fields those numbers depend on are the ones in the Google Business Profile checklist. Underneath all of it sit the numbers a business already tracks without any SEO tool at all — phone calls, form fills, quote requests, booked work.
Does this change the SEO work itself?
No. The work is what it was: pages that answer the question, a profile that is complete and current, citations that agree with each other, a site that loads. What changed is how much of that work is externally observable, and observability is not the same thing as performance.
The harder version of this problem is already here: Answer Engine Optimization has never had a ranked list to scrape, so the habit this change forces — reporting on outcomes you own rather than positions you borrow — is the one that surface needs anyway.
Frequently asked questions
Does the GoTo URL change affect my rankings?
No. It changes how a click on a result is routed, not how results are ordered, so your positions are unaffected by the change itself. What it does affect is the third-party data collection that reports those positions back to you.
Why is my rank tracker showing gaps or a sudden drop?
Most likely because the tool could not read the destination URL for some results and recorded nothing rather than a position. Check whether the gap coincides with your provider's own status notices before treating it as a ranking loss, and compare against Google Search Console for the same dates.
Is Google Search Console data affected?
Search Console is first-party data reported by Google itself, so it does not depend on reading the results page from outside and is not affected the way scraped data is. That makes it the better reference point while third-party tools adjust, even though its position metric is an average rather than a daily rank.
What should I tell a client whose ranking report looks wrong this month?
Tell them before they ask, and tell them it is a measurement change rather than a ranking change, naming the redirect as the cause. Then show the first-party numbers that are unaffected — Search Console clicks and impressions, calls and form fills, Business Profile actions — so the conversation moves to outcomes that were never scraped in the first place.
Does this change what SEO work I should be doing?
Not the work itself. It changes what you can observe about the work, which is an argument for reporting on first-party outcomes rather than on borrowed position data, and for treating rank grids as one input among several.