AI Search Isn't Hitting Publishers Evenly — and the Gap Should Change Where You Place Content

The interesting thing about AI answers eating search traffic isn't that it's happening. It's that it isn't happening equally to everyone, and the shape of the gap tells you something useful about where to put your work. If you want the practical takeaway up front, it is this: place content where it can earn attention beyond search.

Published analysis puts organic traffic loss at roughly 60% for sites with between one and ten thousand daily views, compared with roughly 22% for sites with over a hundred thousand daily views. The studies vary, and I'd treat any single figure with suspicion, but the direction is consistent across all of them: the smaller you are, the harder it lands.

I'm James Nash. I run Home Hustle Hub, a US side hustle and productivity site, on my own alongside a full-time job, which puts me squarely in the group the data says loses the most.

The reason for the asymmetry is straightforward once you see it. Large publishers have brand demand. People search for them by name, arrive directly, and would have done so with or without a blue link. A small site usually had exactly one asset: it ranked for a question somebody typed. When the answer engine answers that question above, the asset is gone, and there is no second route in.

This is starting to be argued rather than absorbed. In June 2025, the Independent Publishers Alliance filed an antitrust complaint with the European Commission over AI Overviews, backed by the Movement for an Open Web and Foxglove; the same coalition later took the argument to the US Department of Justice. In the US, the Journalism Competition and Preservation Act proposed a temporary safe harbor for news publishers to collectively bargain with platforms rather than one-on-one.

The precedent I keep coming back to is older. ASCAP was founded in 1914 because individual songwriters had no practical way to license or collect on every performance of their work. The structural problem is the same one: value taken in very small increments across an enormous number of uses, by an entity far too large for any single creator to negotiate with. Nobody solved that by writing better songs.

For marketers, the practical consequence is a change in how you value a placement. A guest article or expert contribution on a small, well-targeted niche site used to be worth doing for the qualified traffic it sent, almost regardless of the host's size. That calculation has moved. If the host is in the band losing sixty per cent of its search arrivals, you are buying into a declining audience, and the relevance of the match no longer compensates.

What I changed on my own site follows the same logic. A page has to be worth arriving at from somewhere that isn't a search result — a citation, a newsletter, a mention inside someone else's article. I stopped treating a ranking as the finish line because it's now a position on a results page that increasingly answers the question before anyone scrolls. That's a different brief for a writer than "target this keyword," and it produces different pages.

None of this is temporary, and none of it is a plateau. Brand demand compounds, and its absence does too. The publishers who come out of this intact will be the ones worth looking for by name. The harder job now is becoming one worth finding.