Google Is Making Search Results Harder to Measure

For years, businesses could ask a simple question: Where do we show up on Google?

A piece of software would check. It would load Google's results, read the page and report that a company ranked third for one search, ninth for another and nowhere for a third.

That system is starting to break. Since around September 13, 2026, Google has become more successful at blocking the automated tools that read its search results, according to Search Engine Roundtable. One rank-tracking company reported losing about 80 percent of the data it had been able to collect.

The problem arrives just as marketers are trying to measure something newer: whether their businesses appear inside answers written by artificial intelligence.

What Is Answer Engine Optimization?

Answer Engine Optimization, or AEO, is the practice of making a business, product or piece of writing more likely to be named and cited when an AI system such as ChatGPT, Google's Gemini, Google's AI Overviews, Perplexity or Claude answers a question.

It follows an older idea, search engine optimization, or SEO. SEO focused on getting a webpage to rank higher in a list of search results. AEO asks a different question: when a machine writes the answer itself, does it mention you?

Google now has several places where that can happen.

AI Overviews are the AI-written summaries Google places above its normal search results. AI Mode is Google's chat-style search that answers questions in conversation.

That creates a measurement problem. A business may want to know where its website ranks in Google's regular results, whether it appears in an AI Overview and whether Google mentions it in AI Mode.

Much of the software built to answer those questions depends on reading Google's results automatically.

And Google is making that harder.

What Happened to the Rank Trackers?

Rank tracking is software that checks, usually every day, where a website appears in Google's results for a list of search terms.

To do that at scale, these companies often use scraping, which means having software load pages automatically and copy what is on them.

Around September 13, Google became more successful at blocking AI scrapers, search scrapers and third-party tracking tools, Barry Schwartz reported in Search Engine Roundtable on September 18.

Derek Perkins of Nozzle, a rank-tracking company, said he saw "about an 80% drop" in the data he could get from Google.

Sistrix, a German SEO-tools company, began collecting data at a reduced rate on September 16. Its status notice said Google "has made further changes to the way search results are delivered, and as a result, data collection is currently running at a reduced rate."

Steve Paine of Sistrix said Google had been changing the delivery of search results for about a year and was now doing so "at a very high pace."

The effects were uneven. Search consultant Glenn Gabe said Ahrefs was still scraping successfully, Semrush was still working but more slowly, and Sistrix had fallen far behind. DataForSEO reported similar trouble.

Schwartz warned that data from these tools "might be wrong" during this period.

The article included no statement from Google about the September blocking.

Why Does Google Keep Changing Access to Its Results?

The September disruption was not the first change.

In the second week of September 2025, Google stopped honoring a URL parameter called "&num=100," according to Search Engine Land. The parameter had allowed rank trackers to request 100 search results on one page.

Without it, collecting the same amount of information required ten times as many requests. Semrush and Accuranker both acknowledged disruptions.

A Google spokesperson told Search Engine Land that use of the parameter was not something Google formally supported.

Then came another change in 2026.

Beginning in July, Google tested a passthrough address called google.com/goto. By late August, it had rolled out broadly. Instead of sending someone directly from a Google result to a website, the click first passed through a Google address.

Perkins of Nozzle reported "nearly a 100% rollout across several residential ip providers."

Google confirmed the change. "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," the company said.

Schwartz wrote that the change appeared aimed at blocking scraping by third-party tools and AI companies. By September 1, he reported that the tools had already developed fixes.

Then came the September blocking.

Three changes in roughly a year had the same practical effect: outside software had to work harder to read Google.

Why Is Google Fighting Scrapers?

The dispute is also playing out in court.

On December 19, 2025, Google sued SerpApi, a company that sells access to Google results through an API.

Google general counsel Halimah DeLaine Prado wrote that SerpApi used "shady back doors" and large networks of bots to get around Google's security measures. Google's complaint said SerpApi sent "hundreds of millions of artificial search requests each day."

SerpApi's general counsel, Chad Anson, described the service differently. "The information we provide is the same information any person can see in their browser without signing in," he said.

In July 2026, Chief U.S. District Judge Yvonne Gonzalez Rogers dismissed Google's suit.

She wrote that "plain and aggregated search results such as URLs, snippets, and factual index data, are publicly accessible facts and are not 'works protected under the Copyright Act.'"

Claims based on results containing no copyrighted material were dismissed without leave to amend. Google received 21 days to refile a narrower claim involving licensed images in Knowledge Panels.

The ruling addressed copyright. It did not stop Google from changing the technical systems that automated tools must get through.

Google's own statement on the goto change said the company regularly deploys measures against what it considers abuse.

Is Search Itself Becoming Harder for Machines to Read?

Google is not alone in putting barriers in front of automated access.

On September 18, Search Engine Roundtable reported that Bing was asking some searchers to click a box to "verify you are human" before showing results. The report was based on a screenshot from Sachin Patel. Schwartz wrote that Google was doing similar things "in various forms."

At the same time, Google's search interface is becoming less centered on the old list of links.

Google was testing an AI Mode button inside the search bar on results pages, Search Engine Roundtable reported on September 18. Dwayne Cubbins, who spotted the test on September 15, asked whether Google was trying to push people toward AI Mode when they made a follow-up search.

On September 20, another Google experiment showed how fluid the interface has become. A bug in Web Guide, an opt-in interface combining AI answers and web results, left some searchers unable to return to classic results. Gabe said he was sometimes "stuck in Web Guide and can't get out."

For measurement companies, every new format creates another thing to detect, collect and interpret.

Why Does This Matter to Publishers and Businesses?

Referral traffic means visitors a site receives from another site, such as Google.

Publishers say that traffic is already falling.

At closed-door sessions at the Digiday Publishing Summit, publishers described Google referral declines of 15 to 40 percent during the past year, according to a September 17 report by Sara Guaglione. Forbes reported a 40 percent year-over-year decline.

One publishing executive said, "The days of making money on mass traffic from Google referrals are generally down and over."

Esther Cohen of The Verge said she now thinks of Google traffic as extra and does not like to plan or optimize around it.

Publishers at the summit cited click-through rates of about 1 percent from AI Overviews.

But one publisher also reported a different result. Work aimed at optimizing content for AI systems produced a 40 percent increase in customer acquisition and measurable improvements in being surfaced by ChatGPT.

That helps explain the interest in AEO. Businesses still want to know where they appear. The place they appear is changing.

What Can a Business Still Measure?

The weak point is now clear.

The measurement layer of SEO and AEO depends heavily on machines reading Google's results. Google has narrowed that access three times in a year: the removal of num=100, the goto redirect and the September blocking.

That does not mean every ranking report is useless. It means some reports this month may be incomplete.

There are still measurements that do not depend on outside companies scraping Google.

Google's own Search Console can show a site's performance in Google search. A site's analytics can show referral traffic, including visits arriving from ChatGPT, Perplexity and Gemini. And a business can inspect the answers visible inside AI systems by asking the systems directly.

Those methods measure different things. None provides one perfect score.

That may be the central lesson of AEO so far. Search is moving from a page of links toward systems that write answers. At the same time, the outside tools built to observe those systems are losing access to some of the data they used to collect.

Businesses still need to know whether they are being found.

For now, they may have to answer that question with several imperfect views instead of one dashboard.