Part 1 - Too Small to Sue, Too Small to License, Ingested all the Same
Part 2 - The Answer Without the Author
Part 3 - The Machinery of "No"
This is the second post in our series. For twenty-five years or so, the web ran on an implicit bargain. Publishers allowed crawlers to index their content freely, and in return search engines sent readers back to them as traffic. Nobody ever wrote the arrangement down, but it held because both sides genuinely needed it. AI answer engines have now ended that bargain, and they have done so more or less unilaterally.
What the data shows
The data on this is fairly stark. Similarweb measured zero-click searches (queries that end without a single click through to any website) at 69% by May 2025, and by 2026 that figure has climbed above 80%. A Pew Research Center study of some 68,000 real queries found that when an AI summary appears in the results, users click through to an organic result only 8% of the time, against 15% when no summary is present; only 1% click the citation links inside the summary itself. Ahrefs, meanwhile, looked at the click-through rate for content ranked first on informational queries (precisely the kind of query where the underlying knowledge matters most) and found it had fallen from 7.6% to 1.6%. That is a collapse of 79%, and for content sitting in the top position.
The other side of the ledger is just as telling. Cloudflare's data puts Anthropic's crawlers at roughly 70,000 page requests for every referral visit they send back. The pattern, in other words, is one of taking at scale and returning almost nothing.
A revenue crisis for some
For news sites, travel guides and general-interest media, this is straightforwardly a revenue crisis, and some have already closed their doors entirely. "The extinction-level event is already here," as Helen Havlak, publisher of The Verge, has put it.
For researchers, the threat is different
For researchers, NGOs and archives, the threat is a different one, and in some ways a sharper one. Their revenue doesn't primarily depend on web traffic in the first place. Their credibility does. Citation and attribution are the currency that academic and mission-driven work runs on. They are what justifies the grant, validates the methodology, and establishes the institution. The point of publishing a study, a policy paper or a dataset is not only to inform; it is also, and just as importantly, to be seen as the source that informed.
And so an AI answer that draws on a body of research without naming it (or that lists it only in citations which 99% of readers never click) does rather more than lose a visit. It breaks the chain of attribution that makes the work legible as knowledge rather than as so much background noise. A paraphrase with no name attached makes an institution disappear in plain sight, even as its work is actively being used.
What it means to be a canonical source
The question that follows from this isn't only how we recover the lost traffic. It is a larger one. In a world where AI systems have become the primary interface between knowledge and the reader, how does original work establish that it exists at all, that it can be trusted, and that it came from somewhere specific? Those, it seems to us, are not really SEO questions. They are questions about what it means to be a canonical source.
Next in this series
Next in this series, we'll look at what the standards bodies are actually building in response, and at how far along they really are.