The death-of-the-web pieces argue for building one

Mike King published *Machine Media: The Death of the Open Web* on Tuesday. Rand Fishkin published *Inimitable Product is the New "Make Great Content"* the day before. Read them back to back and you get the impression the web is in the late stages of a hospice stay, with Google and OpenAI standing over the bed deciding which organs to harvest first.

Both pieces are well-argued. Both contain claims I'd defend in a meeting. But the conclusion most readers are taking from them — that owned media is dead, that the website is finished, that the only path forward is to either fight the monopolists in court or invent ultrasonic chef's knives — is exactly the wrong read.

The honest reading of King and Fishkin together is the opposite. The case for owned media has not been weaker in fifteen years. It has also never been more strategically valuable. Both things are true at once, and the gap between them is where most consultants are about to make the wrong call on behalf of their clients.

What King and Fishkin actually said

Let's not strawman either of them, because both pieces deserve to be engaged with seriously.

King's argument is that the internet's primary interface is shifting from browsers and feeds to agents — ChatGPT, NotebookLM, Copilots, ambient assistants — and that the data backs this up. AI bot traffic on Cloudflare's network grew 187% in 2025 while human traffic grew 3.1%. ClaudeBot crawls roughly 24,000 pages for every referral it sends back. GPTBot sits at 1,276 to 1. Global publisher Google traffic fell 33% in 2025. Small publishers lost 60% of their referral traffic in two years; some lost 90%. His point isn't that websites disappear. It's that they become data sources for agents that mediate everything between the brand and the customer.

Fishkin's argument is sharper and more political. Google is no longer trying to be the gateway to the web — it's trying to *be* the web, using its index to power AI answers that make the original content commercially redundant. His prescription has two parts: long-term collective action through politics, and short-term survival through building inimitable products that can't be summarised away.

Both writers are correct about the structural shift. Where I diverge from the readership of both pieces — not necessarily from the writers themselves — is on what to do about it.

The collapse argument has a missing premise

King's framing is that owned media is no longer the destination. That is true if you measure destination as "place where humans land via search referrals." It is not true if you measure destination as "source of record that machines and humans both consult."

The death-of-the-web argument is really an argument about referral traffic, dressed up as an argument about owned media.

The missing premise in the death-of-the-web argument is this: agents need somewhere to ground their answers. They cannot hallucinate brands, products, prices, opening hours, or policies without consequences. When they do, users lose trust in the agent, not the brand. So the agents need authoritative sources, and the structure of that need has not changed. They need entities. They need verifiable facts. They need fresh inventory data. They need consistent representations of who you are and what you sell across the surfaces they can read.

That sounds like a website. It is, in fact, a website. Just one that's optimised for machine consumption rather than human session time.

The shift King is describing is real. But the conclusion isn't "websites are over." It's "the value of your website now flows through agents rather than through human visits to it." Those are radically different statements. The first sells panic. The second sells a roadmap.

The death-of-the-web argument is really an argument about referral traffic, dressed up as an argument about owned media.

Fishkin's prescription is right. The framing makes it sound impossible.

Build inimitable products. Yes. Of course. Anyone who's done this work for more than five years has been giving versions of that advice since long before AI Overviews existed. The businesses that compounded — the ones that didn't get flattened by every algorithm update — were the ones that had something proprietary at the centre. A real product. A real service. A real brand people sought out by name.

What I'd push back on is the implication that "inimitable" requires the level of craft Fishkin describes — ultrasonic chef's knives, made-to-measure suits with whimsical personality, pottery techniques refined over a thousand years. Those are wonderful examples. They are also examples that 99% of the businesses reading his post will never apply, because they sell accounting software or run a roofing company in Sheffield or do dental implants in a market town.

The honest version of "build something inimitable" for most businesses is much smaller and more achievable:

• Have actual expertise that shows up in your content
• Maintain a real opinion on your subject matter
• Provide service no commodity competitor will provide
• Build genuine relationships with the people in your market
• Be findable by name because people in your industry know you exist

That's not building a thousand-year pottery tradition. That's just running a competent business with a coherent identity. And the place that identity lives — the source of record for every claim a machine will make about you — is your website. Owned media. The thing both pieces have written off as past tense.

The actual position both pieces support, even if they don't say it

If you sat King and Fishkin down separately and asked them whether businesses should give up on their websites tomorrow, neither would say yes. King would tell you to restructure your site as a data source for agents — feeds, schema, MCP endpoints, structured APIs, the lot. Fishkin would tell you to build a brand strong enough that people search for you by name even when AI tries to steer them to a generic answer.

Both of those positions require owned media. They require it more than the last ten years did, not less.

The reason readers are taking a different message away is the rhetorical framing. "Death of the open web" and "the great digital enclosure of publishing" are punchy. They go viral. They get linked. But the operational takeaway from each piece, if you actually read past the headline, is closer to: *the website you have now isn't the website you need, but you still need a website.*

That's a much less shareable framing. It's also the accurate one.

What changes for businesses planning their next twelve months

I've had three separate conversations in the last fortnight with marketing managers who have read pieces like King's and Fishkin's and arrived at the same conclusion: maybe we should stop investing in the site and pour everything into LinkedIn, or YouTube, or paid acquisition, or whatever surface they think the audience is migrating to.

This is the wrong response, and it's predictable. When a high-status writer publishes a piece declaring something dead, the discount-rate on that thing drops in everyone's head. Investment follows perception. Six months later the businesses that internalised the panic find themselves with a thin owned-media asset and a dependency on rented surfaces that the same writers will be declaring dead a year from now.

Here's what actually changes, in my experience, for a UK business planning the next year:

The proportion of effort that goes into *human-facing* page polish — homepage hero copy, prettier service pages, blog posts written to entertain rather than inform — should shrink. Not to zero. But shrink.

The proportion that goes into *machine-facing* clarity should rise sharply. Schema that's actually maintained. Entity consistency across your site, your Google Business Profile, your Wikidata entry if you have one, your LinkedIn, your industry directory listings. Product feeds, where relevant. APIs and MCP servers, if you have the technical capacity. A site structure that an agent can navigate as easily as a human can.

The proportion that goes into *brand* — being a thing people search for by name, get recommended by name, mention in Reddit threads by name — should rise even more sharply. This is the lesson the SEO industry keeps refusing to learn. Branded search is the highest-converting traffic any site gets, and it's the most defensible against AI summarisation because it represents intent that already names you.

None of that means your website is dead. It means your website is doing different work for a different reader.

The honest limits

The bit of King's argument I can't easily refute is the agentic transaction layer. If the agent doesn't just recommend but also transacts — checks out, books, contracts on the user's behalf — then the brand really does become a fragment inside someone else's experience. That's a meaningful shift, and I don't think the SEO industry has reckoned with it.

But agentic transactions aren't here yet at any meaningful scale for most UK businesses. ChatGPT can recommend a plumber. It can't book one. Perplexity can show you holiday options. It isn't taking your credit card. The infrastructure King describes will arrive — possibly faster than I expect — but the gap between announced capability and production deployment is wider than the discourse implies. Businesses that retreat from owned media in 2026 on the assumption that 2028's infrastructure is already operational are going to spend three years invisible for no good reason.

The bit of Fishkin's argument I'd partially concede is the regulatory pessimism. He's almost certainly right that political action won't save the open web on a fast enough timeline to matter for individual businesses making decisions this year. I'd just add that *no* solution arrives on the timeline of individual business decisions. You can't wait. You build the inimitable thing now, you publish it on the surface you control, and you accept that the discovery layer above it is volatile.

What I'd actually tell a client this week

If you came to me on Monday saying you'd read King and Fishkin and were ready to give up on the website, here's what I'd say.

Don't. Your website is more important than it was three years ago, not less — but it's important in a different way. It used to be where humans landed. Now it's where machines verify what humans were told about you. Both roles matter. The second one is the one your competitors are mishandling, and that's your opening.

Invest in the parts of the site that make you machine-legible: schema, entity clarity, structured data, accurate inventory, technical performance, semantic HTML. Invest in the parts that make you human-meaningful: a real point of view, real expertise, content nobody else in your category produces because they don't have the experience to produce it. And invest in being known by name in the communities — online and offline — where your buyers actually spend time.

That's not death-of-the-web strategy. It's just strategy. The fact that it sounds boring next to "the great digital enclosure of publishing" is part of why nobody's doing it well.

The web isn't dying. It's becoming less forgiving to businesses that treated it as a traffic-acquisition surface and more rewarding to those that treat it as the place their identity lives. Both King and Fishkin would agree with that read of their own work, I suspect, if you took the headline away.

That's the loop. The doomsayers and the practitioners are saying the same thing. The doomsayers just get more retweets.