Back in antiquity (five years ago), when I ran a popular Web 1.0 content site called Beliefnet, we used to cockily predict to investors that our advertising rates were going to rise every year. We knew this because the prestigious market researchers told us so. And their logic seemed flawless: More and more ad dollars were going to shift online, and improved ad targeting technology would improve CPMs.

We all know why the first trend didn’t lead to a windfall for publishers (the money did move online, but most of it went to Google, Facebook, and Yahoo). But why haven’t ad tech improvements provide a windfall for most content publishers?

The answer is a revolution in how advertisers view the importance of what content is surrounding their ads. Many advertisers now care more about who sees their ads than where they appear. Context no longer matters so much.

That poses huge problems for publishers who invest in costly forms of content creation (sometimes known as “reporting”)—and it partly explains the escalating acceptance of native advertising. Debates have intensified about the pros and cons of native—a study this month showed that native risks harming publishers’ brands—but it’s worth understanding how publishers got into this position.

For most of modern media history, advertisers spent with a media outlet for two reasons: to reach a certain type of person and to have their brand rub up against the publication’s brand.

For the first goal, it was an inexact science. If your pimple cream wanted to reach, say, 17-year-old teenage girls, you’d advertise in Mademoiselle. If your local hardware store wanted to reach middle-age guys in town, you advertised in the newspaper. Truth is, it was pretty inefficient for advertisers, since they were also paying for the many other readers of the newspapers who didn’t care much about lawn care. But it was the only option advertisers had.

The publications could tell themselves that the advertisers wanted to be alongside the great content, but really advertisers were mostly viewing context as a proxy for reach.

Of course the internet blew up that formula: Advertisers can target their ads to particular demographics with precision.

At first, it seemed this targeting might be fine for publishers. For instance, in the early days of Beliefnet, we could charge pharmaceutical advertisers more for the health section than the religion section of the site on the assumption that the readers of health content were more sickly. And we could target health-content readers even as they went to other parts of the site. Sure, it created perverse incentives for us to produce twice as much health content as religion content, which was awkward since we were, well, a religion website. But it worked for a while.

Then technology evolved further in a way that proved harmful to publishers.

Google got better and better at providing advertisers the audience they wanted based on search results, obviating the need for many advertisers to place banners on websites. Facebook did the same.

Then came the birth of programmatic ad buying, which meant that advertisers could reach their precise target audience with little regard to where the ads appeared. Using a variety of technology, programmatic buying enables advertisers to find the cheapest, most cost-effective inventory, including with “real-time bidding.” Now, an advertiser goes to an agency and says, I’d like the best possible rate for reaching 25-year-old single men, and the agency will provide them with inventory throughout the entire internet. The key is getting the best possible rate. They can basically scour the Web—where volume continues to grow rapidly—for the cheapest way to reach that type of person.

That’s not to say that advertisers don’t want to reach certain types of readers. But a marketer can now reach “New York Times readers” without ever actually advertising in The New York Times, and for less money than sending a check to the Gray Lady.

The result is that all this great new targeting technology has put downward pressure on ad rates for publishers.

This is all a bit strange for anyone who experienced the old ad system in which advertisers were hyper sensitive about what content was going to be adjacent to their ad. Newsmagazines had to be careful not to put upbeat ads next to disaster coverage. At Beliefnet, we had Christian advertisers request that their ads not appear in the Wiccan section. (We used to joke that we should charge them more to appear in non-Christian areas since they might rack up some religious conversions over there).

Now, all sorts of well-preserved brands have their ads appearing in all sorts of second-rate places. And they don’t seem to care! Many advertisers don’t know where their ads appear.

Steven Waldman was senior advisor to the Chairman of the FCC and principal author of its report on the changing media landscape. He was chair of the Council on Foundations Working Group on Nonprofit Media and is a consultant to the Pew Research Center. Before that, he was the founder of and a national correspondent for Newsweek.