Why 69% of Google Searches Now Skip Your Website Entirely

In May 2023, Google quietly launched something called Search Generative Experience. It appeared in a small percentage of searches, mostly for testers. Most publishers, bloggers, and digital marketers looked at it and moved on. It wasn’t their problem yet.

By May 2024, it was renamed AI Overviews and rolled out to every US user. By early 2025, it was appearing in over half of all Google searches. And somewhere in that 18-month window, a business model that had sustained tens of thousands of websites, bloggers, affiliate marketers, and publishers for two decades started quietly coming apart.

Not with a dramatic announcement. Not with an algorithm update you could see in your Search Console. Just a slow, steady draining of the traffic that everything was built on.

That number deserves a moment. In May 2024, 56% of searches ended without a click. By May 2025, it was 69%. In twelve months, Google retained 13 more percentage points of user attention inside its own walls. Every one of those retained users is a visitor someone’s website didn’t get.

Pew Research tracked 68,879 real Google searches in March 2025, not estimates, actual browsing behavior from 900 US adults. When an AI Overview was present, users clicked on search results 8% of the time. Without an AI Overview, that number was 15%. A 47% relative reduction in clicks.

And here’s the part that makes the citation numbers almost darkly funny: only 1% of users click the source links that appear inside the AI Overview itself. Google cites you, tells the world you were the source, and 99% of the people who saw that citation never visit your page.

Wikipedia is the most cited domain in Google’s entire AI system with over 1.1 million mentions, and it still saw an 8% decline in human pageviews year over year. The Wikimedia Foundation attributed it directly to AI search. If Wikipedia can’t convert citations into traffic, the picture for a mid-tier blogger writing recipe content or product reviews is not encouraging.

This didn’t happen overnight, and the pace of the shift matters because it tells you something about where the next two years go.

Answer Engine Optimization. You’re going to hear this term constantly over the next two years. Here’s a grounded version of what it actually means.

Search Engine Optimization was the practice of getting your page ranked highly on a list of links. The click was the product. Your ranking determined how many you got.

Answer Engine Optimization is the practice of getting your content quoted inside the answer itself. When someone asks ChatGPT, Perplexity, or Google AI a question, there is no list of ten links. There is one synthesized response citing 2 or 3 sources. AEO is how you become one of those sources.

The tactics build on what SEO practitioners already know but push in new directions. Content needs to be structured around explicit questions and direct answers. Pages need regular updates. Content refreshed within 2 months earns 28% more AI citations than older pages. FAQ sections correlate with higher citation rates. Long-form content with clear sectioning outperforms thin pages.

But the underlying requirement is genuine authority. AI systems have been trained on enormous amounts of web content and they’ve learned, imperfectly but meaningfully, which sources are credible. You cannot schema-markup your way to AI citations if your content doesn’t have real depth behind it.

Everyone in digital marketing is getting broadly the same advice: create quality content, build genuine authority, optimize for AEO, diversify your traffic sources. That advice is correct. It is also the same prescription for two completely different economic situations.

The blogger and the brand are not experiencing the same crisis. They are experiencing the same technology creating opposite outcomes.

Their revenue is the click. Ad impressions, affiliate commissions, page RPM, every monetization model in this category requires a human body to physically land on a page. When AI answers a question before anyone navigates to a website, the content was used as raw material and the economic benefit stayed inside Google’s ecosystem.

Affiliate revenue is down 50-70% across mid-tier publishers. One veteran publisher tracked by WebmasterWorld went from $500,000 per year in AdSense to $70,000. Business Insider lost 55% of its organic traffic and cut 21% of its staff. Google’s AdSense now represents just 10% of Google’s total advertising revenue, a historic low. Google kept 90% for its own properties.

When AI describes what a brand’s product does, who it’s for, what problem it solves, how it compares. That is a commercial. A free commercial, running inside a channel that reaches 1.5 billion monthly users through Google AI Overviews alone.

Brands have always understood something publishers are only now being forced to reckon with: you don’t need a website visit to move someone toward a purchase. Television advertising never sent you to a URL. A billboard never tracked a click. AI search is building brand recall at scale for companies with genuine authority, whether they asked for it or not.

And when those users do eventually click through, they arrive as much better buyers. AI-referred traffic converts at 4 to 5 times the rate of traditional organic search. These users have already had a detailed conversation about their problem. They arrive with intent that organic traffic rarely has.

The same AI mention that quietly destroys a blogger’s income stream is free advertising for a brand. Same technology, same citation, same mechanism. The outcome is opposite because the economic relationship with the visitor is opposite.

The advice splits cleanly depending on what your economic relationship with the visitor actually is. Get that diagnosis right first, because the strategies point in different directions.

One thing applies to everyone: stop publishing the same article on 20 platforms thinking that multiplies your reach. It splits your authority. Google picks one version to index and credit. If LinkedIn or Medium has more domain authority than your own site, and they often do, they rank and you spent your energy building their platform, not yours.

The competitive dynamic in AI search is meaningfully different from traditional SEO. Google showed ten links. AI shows 2 or 3 sources. There is no page 2. The difference between being cited and not cited is total. This makes early authority establishment significantly more valuable, and late entry significantly more expensive.

The open web was built on a bargain: create content, earn traffic, monetize attention. That bargain is being renegotiated, with one party at the table and the other watching the terms change in real time.

For some businesses, the shift in AI search is genuinely an opportunity. For others, the bloggers who built something real over years of work, the affiliate marketers who learned a complex craft, the numbers are already moving in the wrong direction and the window to adapt is measured in months, not years.

Both groups are being told the same things. They need different answers.

The same disruption reshaping search is reshaping every layer of how organizations operate. AI isn’t just changing where your traffic comes from. It’s changing what skills matter, what roles survive, and what kind of leadership is actually required when the tools keep moving underneath you.

If you work with teams navigating this, product managers are sitting at the center of the AI shift in ways most organizations haven’t figured out yet. And the burnout showing up on your team right now probably has less to do with AI than you think.

A lot of what I write about comes back to the same core problem: the gap between how organizations say they operate and how they actually do. If that resonates, Beyond Management is the longer version of that argument, built from boot camps, corporate meeting rooms, and the kind of lessons that only come from being in the room when things go wrong.

The full archive is at ericbevill.com/articles if you want more.