Google adopts google.com/goto URL in search results and complicates SERP scraping

Google confirmed to Search Engine Roundtable that it is rolling out the use of a pass-through URL, google.com/goto, within search results. In practice, when hovering over an organic result, instead of seeing the direct link to the destination domain, the user now sees a Google URL that redirects to the site. The feature had been in testing since July 2026 and now appears in several browsers.

We have a long history of deploying technical measures against evolving forms of abuse, and we regularly take steps to protect our services and users.

-- Google spokesperson, to Search Engine Roundtable

Despite the generic tone, the target is quite clear to anyone who works with search data: making automated SERP scraping harder for third-party tools and AI companies.

What changes under the hood

Until now, the link on an organic result pointed directly to the destination page's URL, rendered in the SERP's HTML. A scraper could read the HTML, extract the href of each result and reconstruct the ranking without following a single link. That's how rank trackers and data collectors have operated for years.

With google.com/goto, the link exposed in the HTML is no longer the final destination and becomes a server-side redirect controlled by Google. According to Derek Perkins, of Nozzle, quoted in the report, Google "has been testing anti-bot measures that eliminate all client-side links in favor of google.com/goto server-side redirects".

The technical point he raises is where it hurts:

Goto links cannot be decoded, so providers will have to follow the redirects. This works in small-scale tests, but Google will likely make this difficult, since every SERP has hundreds of links to decode.

-- Derek Perkins, Nozzle

In other words: before, it was enough to read the HTML once per query. Now, to find out which domain is behind each position, the collector would need to fire an additional HTTP request for each link and see where the redirect ends up. A results page with dozens or hundreds of links turns into dozens or hundreds of extra requests per tracked query. It's a brutal multiplier of cost, latency, and the chance of being blocked by IP.

Perkins says he has seen the behavior at "nearly 100% rollout across several residential IP providers", with a chart showing the jump in the proportion of SERPs that use goto exclusively over the roughly four months of testing.

Why Google is doing this now

The context is not accidental. Roundtable links the change to Google's legal dispute with SerpApi, a results-scraping company that, according to the report, Google sued and has been struggling to win against in court. If it can't be blocked through legal means, it gets blocked through engineering: making scraping expensive enough to be unfeasible at scale.

The second target is AI companies that scrape the SERP to feed models and answer products. This is where the GEO reading comes in: Google is protecting the index that underpins its own AI Overviews and AI Mode while trying to cut off competitors' cheap access to the same material. It's a data-war move, not a search-quality one.

It's worth noting the skepticism of the report's own author, Barry Schwartz: third-party tools "will adapt; they always adapt". Historically, every anti-bot layer from Google has been worked around. The difference this time is the structural cost imposed, not a simple parsing obstacle.

What this means for SEO professionals in Brazil

Here it's necessary to separate two effects of different natures.

| Aspect | Before | With google.com/goto | |---|---|---| | Link in the SERP's HTML | Direct URL to the destination | Server-side redirect from Google | | Cost to collect ranking | 1 request per query | 1 + N requests (one per link) | | Who feels it | | Rank trackers, collectors, AI | | End-user click | Goes to the site | Passes through the goto and goes to the site |

For the end user's click, nothing changes in the experience: they type, click, and land on the site. The point of attention is the referrer. Redirects can alter or mask the origin information passed to the destination server, which, depending on how Google implements it, has the potential to mess up attribution in server logs and in tools that rely on the HTTP referer. The source doesn't detail this behavior, so it remains a point to monitor, not a confirmed fact.

The direct and measurable impact falls on rank-tracking tools. If your SEO process in Brazil depends on a rank tracker (Nozzle, and by extension any collector that reads the SERP), it's reasonable to expect, in the coming weeks:

• Increased provider costs, possibly passed on in pricing.
• Higher latency in position updates.
• Gaps or instability in data while vendors rewrite their collectors.

The practical recommendation, measuring before and after: compare the keyword coverage tracked and the update frequency of your rank-tracking dashboard this week against a month ago. If it drops, it's not a problem with your site, it's goto hitting your provider. And check whether the origin data on your server (via logs or GA4) remains consistent with what Google Search Console reports, since Search Console uses Google's own data and doesn't depend on scraping, remaining the most reliable source for organic position and clicks.

What remains open

Google has not explained implementation details, the timeline for a complete rollout, or whether the goto format will be rotated to make decoding even harder. There's also no word on the effect on referrer or on publisher-side analytics tools. For those who depend on third-party data, the defensible strategy is to diversify measurement sources, prioritize first-party data (Search Console, logs, GA4), and treat rank trackers as a complementary signal, not as the sole truth, while the scraping market recalibrates.