Instead of an organic result linking directly to the page it found, Google is increasingly returning something like google.com/goto?url=.... Clicking it still takes you to the website, but the value in that url parameter is an opaque Google-specific identifier rather than an encoded copy of the destination URL.

Google has wrapped search-result links before, so redirects themselves are not new. The older google.com/url?q= format still exposed the destination in a way that was easy to extract. The new goto format is different. If software wants to discover where the result actually points, it has to make another request to Google and inspect the Location header returned by the redirect.

For a person clicking search results in a browser, very little appears to have changed.

For software trying to use Google as an index of the web, quite a bit has changed.

Autom believes the change is primarily intended to make large-scale scraping of Google results more difficult. Instead of downloading a search-result page and immediately extracting every destination, a scraper now has to go back to Google for each individual result. That makes scraping slower, noisier and considerably easier for Google to identify and rate-limit.

Google has confirmed that the change is intentional, although it has been less specific about the motivation. A Google spokesperson told Search Engine Land that the company has a long history of introducing technical measures against evolving forms of abuse and regularly changes its systems to protect its services and users.

I imagine AI search engines and the growing number of systems harvesting search results are a large part of what Google is reacting to, but Google hasn't actually said that. SEO scraping has existed for a long time too, and companies have built entire businesses around extracting Google's results.

What I find interesting is that Google is making this change now, nearly three decades after building one of the most important companies on the internet by crawling everybody else's websites.

Search used to be a way out

The original magic of Google was how quickly it got me somewhere else.

I typed something into a little box, Google ranked the pages it knew about, and I went exploring. The value was the index and PageRank. Google helped me find another person's website.

That relationship has slowly changed.

For me, one of the important moments was Google's introduction of the Knowledge Graph in 2012. Google described it as moving search beyond matching strings toward understanding "things" and their relationships. It was technically impressive, and a lot of it was genuinely useful.

It was also when I started noticing Google becoming less interested in returning websites and more interested in returning answers.

Google's own announcement talked about giving people summaries directly in the search results and even answering questions they were likely to ask next. A few months later, Google was describing its ideal search engine as something closer to the computer from Star Trek, capable of simply answering whatever question you asked.

At the time that sounded like progress.

Looking back, I think that was roughly when the relationship between Google and the web started changing for me.

Google had been a map.

Increasingly, Google wanted to become the destination.

I miss searching the weird web

I find myself using mainstream search engines less and less now.

Part of that is AI. If I need an answer to a reasonably well-defined question, I can often ask an AI system, follow its sources and continue investigating from there. That is ironic because Google's response to AI competition appears to be becoming even more of an answer engine itself.

But sometimes I don't want an answer.

I want the web.

I want the weird little website somebody built fifteen years ago. I want the forum post from someone who actually owned the obscure piece of hardware I am trying to fix. I want the personal review written by somebody who bought the product instead of a 20-page affiliate article explaining what a toaster is before recommending the same five Amazon links as everybody else.

I want the long tail.

There was a period when searching the web felt genuinely unpredictable. A search might return twenty useful pages or absolutely nothing. You could keep clicking through results and eventually wander into some strange corner of the internet maintained by one person who happened to know everything about whatever obscure subject you were interested in.

That wasn't always efficient.

It was frequently wonderful.

Today, searching for products in particular often feels like moving through layers of ads, affiliate sites, SEO content and increasingly AI-generated summaries before finally reaching somebody with firsthand experience. A 2024 study that followed thousands of product-review searches across Google, Bing and DuckDuckGo found that search engines were struggling badly with exactly this problem. Highly optimized affiliate content tended to rank well even when researchers found it less useful, and the authors concluded that search engines were losing a continuing fight against SEO spam.

That matches my experience.

I don't think Google invented bad websites. People have been trying to manipulate search engines almost as long as search engines have existed. Google has an extraordinarily difficult problem to solve.

But Google also built an enormous advertising business around the page where we are trying to find things.

Those incentives are never going to be completely comfortable together.

Sometimes Yandex feels more like old Google

Oddly enough, when I actually want to dig around the web these days, I sometimes find myself using Yandex.

I don't mistake Yandex for some politically neutral public utility. It is a company operating under its own government, business interests and biases, just like every other large search engine has pressures that shape what it shows.

But purely as a web-search experience, it sometimes reminds me of older Google.

I get more long-tail results. Product searches seem more willing to return small sites and actual discussions instead of endless affiliate pages. Searches sometimes uncover old forums, obscure sites and parts of the web that have effectively disappeared from Google's results.

It will even surface some of the seedier corners of the internet that still exist whether mainstream search engines want to acknowledge them or not. I'm not suggesting people should use illegal streaming sites or anything else simply because a search engine indexes them. I find their presence interesting because it reminds me that the web itself is much larger and messier than the carefully shaped slice most modern search products present to us.

That mess used to be part of searching.

You never quite knew what was out there.

The `goto` change is small, but it feels symbolic

The new Google redirect probably won't affect me when I click a search result tomorrow.

Technically, it is a relatively small anti-scraping measure.

But I find it symbolic of the direction Google Search has been moving for years.

A search engine originally built its value by discovering links across an open web and helping people follow them. Now even the link in the search result is becoming an opaque Google identifier that has to pass back through Google's infrastructure before revealing where it goes.

That gives Google more control over who can extract its results, how frequently they do it and how those results can be reused.

From Google's perspective, that makes perfect sense. AI companies and search-data businesses can otherwise harvest something Google spends an enormous amount of money building.

From the web's perspective, the irony is hard to miss.

Google became Google by crawling an internet where people published documents and linked them together openly.

Now Google would rather other machines not crawl Google.

I understand the business logic.

I am less convinced that I like what the web looks like when every large participant starts building walls around the information it has collected from everyone else.

I think Google jumped the shark a long time ago

I used to love Google.

There was a time when it felt like one of the companies genuinely making the internet more useful. Search, Gmail, Maps, Google Reader, Google Books and a long list of other projects felt ambitious in a way I still miss.

Google's old motto was "Don't be evil." I know it was always just a corporate slogan, but it represented something about the version of Google I wanted to believe in.

I don't really believe in that Google anymore.

For me, the mourning period ended years ago. The Google I loved probably disappeared somewhere around the time search began shifting from helping me find information toward trying to become the place where information itself was consumed.

That transition started roughly 15 years ago. The Knowledge Graph arrived in 2012. Answer boxes expanded. Google increasingly put its own products into results. Shopping, travel, maps, knowledge panels and eventually AI-generated answers occupied more of the space that once belonged primarily to links.

Some of those features are useful. I use some of them myself. I even enjoy using Google's AI tools.

But usefulness and stewardship are different things.

I don't think Google is a particularly good steward of web search anymore.

Maybe Google Search is simply following the inevitable path of a mature advertising business. Maybe Google's incentives changed because the internet changed. Maybe building a useful index of an increasingly hostile, spam-filled and AI-generated web really has become extraordinarily difficult.

Probably all of those things are true.

I still think something important has been lost.

A good search engine should want me to leave.

Its job should be to understand what I am looking for and get me to the person, document or website that has it as quickly as possible.

The moment keeping me on the search engine becomes the objective, it stops being a map of the web and starts becoming another walled garden built on top of it.

Now even the links point through the garden first.

Read Autom's analysis of Google's new goto links.

Read Autom's technical explanation of resolving the new links.

Read Google's 2012 introduction of the Knowledge Graph.

Read Ars Technica's coverage of the research into declining product-search quality.

All notes