The number that matters is RPM: revenue per thousand pageviews. Not per thousand visitors — per thousand views, including the ones that bounce in three seconds.
For English-language technology content, publishers commonly report somewhere between five and twenty dollars per thousand pageviews, with developer and B2B traffic at the higher end and consumer traffic at the lower. Treat every published figure, including these, as an anecdote rather than a benchmark: your own traffic mix is the only dataset that applies to you.
The arithmetic, plainly
Ten thousand pageviews a month at eight dollars RPM is about eighty dollars. A hundred thousand is about eight hundred. That is the whole model, and it is why directories chase volume: there is no premium tier in display advertising.
The place it breaks is the assumption that pageviews scale linearly with listings. They do not. A directory with five thousand entries and no links pointing at individual entries gets the same traffic as one with five hundred, because Google is not indexing pages it has decided are interchangeable. Volume without uniqueness is the most expensive mistake in this category.
Three things that move the number more than the ad network
Geography. US and Western European traffic pays several times what most other regions pay. A site with identical content can have a 4x difference in RPM depending on where the readers are.
Page intent. A page someone lands on to solve a problem earns more than a page they bounce off. This is why the comparison and alternatives pages in this directory matter financially, not just editorially: people who read them are deciding what to buy.
Placement discipline. One leaderboard and one rail unit will usually out-earn four aggressive units, because the four-unit page loads slower and gets read less. There is a reason every serious publisher talks about Core Web Vitals before they talk about fill rate.
Where display ads stop being the answer
Sponsored placements and affiliate links both beat display on the same traffic, because they capture intent instead of interrupting it. A single sponsored category slot at fifty dollars a month is worth more than several thousand pageviews at display rates. That is not an argument for putting ads everywhere — it is an argument for keeping the content good enough that somebody wants the slot.
The honest summary: display advertising on an AI directory is a modest business that works at scale and fails below it, and the scale has to come from pages people actually search for rather than pages that merely exist.