A contextual ad network is a company that collects ad space from many publishers and places each ad by what the page is about, not by who is reading it. A publisher is the owner of the website or app. "Contextual" means using the context: the words, the images and the video on the page. The method itself is explained in contextual advertising.
For years, most online ads followed the person. Third-party cookies, small files that a browser stores for companies other than the site being visited, let advertisers recognise one reader across many sites. Privacy laws and browser makers are now limiting this, and many readers block such tracking. A contextual network does not need to know the reader. It needs only to understand the page.
Say a Hindi personal-finance website publishes an article on how to choose a car loan. Soon after the page goes live, the network's software reads it. It finds the main topics: car loans, interest rates and new cars. It also reads the photo, which shows a family beside a small hatchback, using computer vision, software that recognises what is in an image. It files the page under "auto finance" and "car buying". Later that day, a reader in Indore opens the article. The network offers the ad slot to advertisers who have asked for car-buying pages. A car insurance company wins, and its ad for a new-car policy appears. The insurance company never knew who the reader was. It knew only that a person reading about car loans is likely to buy a car soon.
The same reading protects brands. An airline does not want its ad beside a news story about a plane crash, even though both mention flights. So the network also marks pages that are unsuitable for particular brands, and keeps those brands' ads away from them. Advertisers often confirm this afterwards with ad verification tools.
Contextual networks serve sites of every size. A large digital news publisher earns from them. So does a niche news site about cars, health or cricket, whose pages share a clear topic. Advertisers can buy the matched slots through a demand-side platform (DSP).
Some networks go further. They measure attention, meaning how long people actually look at an ad. Others read videos, so that a brand's ad plays only before videos that suit it. One of the largest contextual ad networks was built largely in India, by Indian founders.