This is about behavioural ads vs contextual ones. So what are each of them?
The mainstay of online advertising is "behavioural advertising" in which ads are placed based on dossiers of your activity and preferences that have been compiled by Big Tech giants and shadowy data-brokers.
In contrast to contextual ads:
That's when publishers sell off the right to advertise to you based on the subject of the article you're reading, your location (based on your IP address) and other metadata, like which browser and OS you're using.
So, the issue (for whoever is running the ad, not the user) is the following:
Advertising on the Washington Post is expensive. "Washington Post reader" is a valuable category unto itself: a lot of blue-chip firms will draw up marketing plans that say, "Make sure we tell Washington Post readers about this product!"
Here's the thing: the companies want to advertise to Washington Post readers, but they don't care about advertising in the Washington Post.
For those who don’t get the ad bid, they still have the info that you are a WaPo reader. Now they can advertise to WaPo readers without having to advertise there! This is obviously a problem for the WaPo, because they want to sell their ad spaces, which is harder for them if firms can target the same people for cheaper somewhere else.
So why is the WaPo (and of course other publications!) still doing that?
Data-brokers make crazy claims for how well their targeting works in "conversions" – that is, turning ads into sales.
Doctorow found a paper, that says contextual ads are just as effective. Plus, they’re friendlier towards the user’s privacy (and thus easier to implement under GDPR)