“The answer lies in vector space models: for semantic SEO there is no other way,” Danny told us
Welcome to the tenth episode of SEO Confidential, the series designed for entrepreneurs and SMBs who want to understand and face the complex challenges of digital marketing with clarity.
After our enlightening conversation with Pete Meyers of Moz, today we have the honor of hosting another leading figure in the international SEO community: Danny Goodwin, Editorial Director of Search Engine Land.
With a career that has seen him at the helm of authoritative publications such as Search Engine Journal and Search Engine Watch, Danny has been a witness to and chronicler of the biggest changes in the world of online search for more than fifteen years.
At a time when Google AI Overviews and LLMs are rewriting the rules of online visibility, his analysis is more valuable than ever.
How can we stay visible in the era of “zero-click“?
What will become of our content, now that artificial intelligence seems to treat the web as a free buffet to pull answers from?
These are the questions every entrepreneur working on the web is asking. I decided to launch this interview series precisely to tackle these doubts, offering different, and sometimes diverging, perspectives to decode the future of search together.
The goal is one and one only: to turn uncertainty into strategy and keep growing your brand.
“Blogs aren’t dead. Blogs built solely to monetize organic traffic are”
Hi Danny, my readers are mostly entrepreneurs and SMB managers who are watching the evolution of artificial intelligence in the SERPs with concern. In a context where users find answers without clicking anymore, many are asking: is it still possible to thrive online without organic traffic?
AI Overviews are accelerating the zero-click era, but it’s not a new phenomenon. Google has been eroding organic traffic for years, through featured snippets, knowledge panels and other features that provide instant answers. What’s really new is that today it can answer far more completely, and across a much wider range of questions, without the user ever leaving Google.
This hits informational content hard, which is receiving fewer and fewer visits: for many basic queries, Google already provides everything that’s needed.
However, navigational and commercial searches – those tied to brands, products or purchase intent – continue to generate traffic.
So the challenge we’re facing is figuring out how to own those access points where the click still exists and is worth more.
Google claims that AI Overviews generate higher-quality clicks, but many analyses point to a sharp drop in CTR and conversions. Who’s right? And why is there no transparency on this data in Search Console?
The two things aren’t mutually exclusive. It’s possible that AIOs bring more qualified traffic, but at the same time it’s a fact that many publishers are losing between 20% and 60% of their organic clicks.
The point is that Google doesn’t provide detailed data on AI Overviews in Search Console, and that fuels suspicion. Some think it doesn’t want to show how much traffic it’s keeping, or that the internal data doesn’t really back up its claims.
The truth? We may never know.
If appearing in AI Overviews is becoming increasingly important, how can you monetize a presence that doesn’t bring direct clicks? I know you mostly deal with content, but my readers – entrepreneurs and professionals who work online – want to understand what concrete, measurable strategies they can adopt to turn those mentions into leads or sales.
Some suggest approaching the issue the way you would brand marketing: monitoring the growth of branded searches, direct traffic to the site and signals of rising demand. The goal becomes measuring the indirect impact of visibility in AIOs, even without trackable clicks.
It’s quite a challenge, because everything is becoming more opaque, as Jono Alderson put it, it’s basically “like doing marketing in the dark”.
But investing in strengthening your brand and in search intent tied to your company name remains one of the few genuinely viable strategies today.
Danny, let me ask you straight: will Google become a chatbot? Or will AI Overviews coexist with AI Mode?
For now the two will coexist. But everything suggests that, over time, AI Mode will become Google’s main interface. According to statements from some executives, this shift should happen by 2033, though it will likely come much sooner.
AI Overviews draw information from online content without paying creators anything, merely citing them while taking traffic away from their sites. In your view, should there be a financial arrangement (even for small publishers) for the use of this content?
Well, as you know, in some cases deals do exist. The New York Times, for example, has struck an agreement with Google. OpenAI has signed similar deals with several publishers.
But if you’re a small publisher, the chances of getting paid are slim. Cloudflare recently proposed a new model: Pay Per Crawl, a system in which AI crawlers would have to pay to access content. It’s an interesting idea, even if it’s unclear whether it can really work. What’s certain, though, is that it’s a first signal that the issue of compensation can no longer be ignored.
Why should a blog owner keep creating content for Google today? There used to be traffic and advertising, at least. Now it feels like working for free for Big G. Aren’t we at risk of impoverishing the web, losing original and critical voices?
It’s a bit like asking: “Why open a restaurant after the Yellow Pages disappeared?”
There are other ways to build an audience: social, video, newsletters, communities, word of mouth. Google still sends traffic, but it’s no longer the only channel nor the most reliable one.
The problem arises when you create a blog for Google, rather than for people or out of passion. Blogs born to share ideas, experiences and knowledge still make sense. Blogs aren’t dead. Blogs built solely to monetize organic traffic are.
If content creators are no longer incentivized to produce, where will artificial intelligence and LLMs get their information from? Isn’t there a risk they’ll end up spreading only old, outdated data?
It’s true: artificial intelligence treats the content available online as if it were freeware. But if the web stopped producing new information, AIs would stop evolving, ending up recycling knowledge that’s already known. For certain static things – the sky is blue, the grass is green – that may even be fine. But for everything else, no.
That said, I don’t think content will disappear entirely. It’s more likely that a substantial share of “me too” content will be wiped out: content produced only to chase trending keywords, with no real value.
An example? The recipe sector, where for years everyone published the same variations of the same recipes, hoping to beat the competition on Google.
How can you capture the AI’s attention and appear authoritative in its eyes? Is the famous E-E-A-T still valid, or has SEO changed completely?
The best answer, in my view, is just one: vector space models. That’s where authority is really at stake today. SEO should have changed long ago, ever since Google moved to a semantic search logic, but many tools and approaches remain tied to a purely lexical idea of search, based on exact words.
As for E-E-A-T, to me it has always been more of a concept than a real operational lever. Its origin, in fact, goes back to the web credibility guidelines of the Stanford project, which already explained twenty years ago what makes content credible in people’s eyes.
A small note for you, the reader:
The Stanford Guidelines for Web Credibility are a set of 10 practical rules developed by the Stanford Persuasive Technology Lab in 1999 to help websites and companies come across as credible in the eyes of users. They are the result of three years of research involving more than 4,500 people.
Here’s the gist in short:
How do you build a website’s credibility? By showing transparency, professionalism and trustworthy content.
Among the key points:
- provide verifiable sources and links to reliable external data;
- clearly show who is behind the site (a real company, address, contact details);
- highlight the expertise and authority of those who write;
- take care of the site’s design and usability;
- avoid errors, intrusive advertising and outdated content.
Why did Danny invite us to dig into all this?
Simple: because the principles identified by Stanford underpin what we now call E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness). Google, to this day, keeps looking for similar credibility signals to decide whether a piece of content deserves trust and therefore visibility.
They may not be laws set in stone, as Goodwin says, but they remain essential concepts for understanding which content the Mountain View giant rewards and favors, even in the age of AI Overviews.
Today survival depends on being remembered, not on being seen
From the interview with Danny Goodwin one unequivocal truth emerges: SEO has turned into something more complex and strategic. We’re not just witnessing a drop in clicks, we’re living through a profound metamorphosis of the very concept of visibility.
The era in which the goal was to “create a blog for Google” is over. As Danny points out, today the real challenge is building a brand and content so authoritative and unique that they become the primary source for the answers generated by artificial intelligence.
Optimization is no longer just for an algorithm that ranks links, but for a system that learns, synthesizes and cites. Let me put it bluntly: if AIs don’t mention your brand, you risk becoming invisible.
Danny talks about vector space models. It sounds complex, insider stuff, but it’s actually easier to grasp than you might think.
When we talk about a vector space model, we mean a new way in which Google and other search engines understand the meaning of words, sentences and online content. Until a few years ago, the search engine relied mostly on literal matches: if a user searched for “cheap restaurant Milan”, Google looked for pages containing exactly those words.
Today, instead (and I’ve been saying this for ages), we talk about semantic search.
Thanks to artificial intelligence, Google translates content into vectors, that is, numerical representations that capture the meaning of a text. This allows the search engine to understand that “affordable place in the Navigli area” means the same thing as “cheap restaurant Milan”, even though the words are different.
For those doing SEO, this means it’s no longer enough to insert precise keywords: you need to write content that is clear, useful and consistent with search intent. I’ve said it to the point of exhaustion: a site’s authority is built by offering real value, not by repeating keywords. And this is where the concept of E-E-A-T (experience, expertise, authoritativeness, trustworthiness) finds concrete ground: Google recognizes quality content precisely through these vector models.
The new SEO, then, speaks the language of authority and specialization. It’s no longer about winning on individual keywords, but about establishing yourself as a point of reference across an entire topic area. This requires a strategic investment in brand building, in creating original content and in multichannel distribution that goes beyond simple Google search.
So the question is no longer “How do I rank first?”, but “How do I become the answer?”
There’s no doubt about it: the future belongs to those who can teach artificial intelligence to recognize their value.
All that’s left for me is to warmly thank Danny Goodwin for the time he gave us and for his invaluable insights.
See you in the next episode of SEO Confidential. The analysis of the future of search continues.
#avantitutta
