“Surely The Big Papers Are Where This Gets Won”
Media lists get ranked by circulation. Mine did — nationals at the top, business press under them, trade titles down at the bottom where I never scrolled.
I never checked that order against anything. Sorting a list by reach felt self-evident, so it went unchallenged in every meeting where it came up.
But two companies counted which outlets AI answers actually cite, and my order came out upside down.
The count covers one month, June 2026, and it runs to 129.3 million citations.
Which Outlets Does The AI Actually Pick?
Trade publications. They were cited 213% more often than mainstream outlets, according to a study Press Ranger and OtterlyAI released on 20 August 2026.
A trade publication is the magazine only people inside one industry read — the logistics weekly, the food-manufacturing title, the thing your customers subscribe to and nobody else has heard of.
Two companies did the counting, and both sell into this space: Press Ranger does press-release distribution and PR software, OtterlyAI does AI visibility tracking. They pooled what they each had.
Here’s the scale they worked from:
- 129.3 million citations in June 2026
- more than 20 million cited URLs
- seven AI platforms, ChatGPT among them
Seven platforms and 129.3 million citations is a lot of counting to sit behind one number, and it’s why a 213% gap this wide doesn’t look like the quirk of any single assistant.
The Citations Land In A Narrow Place
Of the citations that went to publishers holding a licensing deal, five media groups took 69%.
And news as a whole — every news page in the count — is 7.2% of all citations. The rest went to pages that aren’t news at all.
Those two figures don’t share a ruler. One is a share inside a subset of publishers; the other is the size of the whole news slot.
- 69% across five groups tells you how crowded it is among the publishers holding a deal
- 7.2% of everything tells you how much of the room that slot occupies
Which means they point at different decisions. The 69% is about how you reach the publishers sitting outside those five groups. The 7.2% is about the ceiling: if press coverage is the only lever you pull for AI visibility, that’s the size of the room you’re playing in.
It’s a bit like queueing for the good table and then working out that the restaurant seats seven percent of the street.
How Much Of This Skew Can You Move?
Some of the concentration you can’t touch. The licensing deals between AI companies and publishers sit outside your reach.
A licensing deal here means an AI company paying a publisher for permission to use its journalism. OpenAI and others have signed them one publisher at a time.
Publishers with an OpenAI deal averaged 10.2 citations per page on ChatGPT. Publishers without one averaged 6.9 — a gap of 48%.
You can’t lobby your way into somebody else’s contract. I went and looked up which outlets had deals, worked out there was nothing for me to do with the answer, and closed the tab.
So leave the contracts as background, and take the part your job actually touches.
Which part is that? Where you pitch. The trade-title gap holds up whether or not the licensing question ever moves.
What Conditions Produced The 213%?
The release never says how a citation was counted. It states no limitations either, which is why the figure can’t go straight into a target on your own dashboard.
- Both companies have money riding on the conclusion that where you place coverage changes your AI citations
- The deal list (91 agreements across 314 domains) comes from Press Ranger’s own database, and nobody has checked how much of the public record it caught
- The line between a “trade” title and a “mainstream” one isn’t published anywhere
This is an observation, not a cause: publishers with deals were already large and already authoritative, so they arrived holding the conditions that get a page cited in the first place. Maybe the deal does the work; maybe the size does. Separating those would take an experiment nobody has run.
The window is a single month, and the 213% comes out of 16 US industries — trade titles took the majority of news citations in 15 of them. If you’re applying this to a media list in your own market, treat the split as a hypothesis and go and check it on your own numbers.
Although none of that gets you a target figure, it’s still enough to reorder a list on. I can’t find anything that measures whether ranking by circulation works.
The Media List Needs A Second Sort Key
Keep the circulation column. It’s telling you something real about reach.
Next to it, put a column for how often each outlet turns up in AI answers. Adding one more sort key to a list that already exists leaves the rest of the list intact.
Once that column has numbers in it, push the trade titles up the order and watch what happens. If the count above holds, I’d expect the same PR effort to buy more citations than it did before.
Then measure it on your own name, counting how often the brand shows up in AI answers after a trade placement lands and comparing that with what you counted before. A flat line means the reorder didn’t do anything.
The wider question of leaning on third-party coverage has its numbers in the piece on the 84% figure. This one narrows it down to which third parties.
One column. That’s the whole change.
Sources
- Press Ranger / OtterlyAI, “Press Ranger and OtterlyAI Release Study Showing Publishers With OpenAI Deals Earn 48% More AI Citations on ChatGPT”, GlobeNewswire, 20 August 2026 (limitations: both publishers hold a commercial interest in the finding — Press Ranger sells press-release distribution and PR software, OtterlyAI sells AI visibility tracking. The release states no limitations and does not describe how a citation was counted. The licensing list of 91 agreements across 314 domains comes from Press Ranger’s own database, and its coverage of the public record is unverified. The boundary between trade and mainstream outlets is not published. The window is June 2026 only, and the trade-versus-mainstream comparison is drawn from 16 US industries, where trade and niche outlets carried the majority of news citations in 15)
