chief marketing officer of Belkin International Inc., noticed an odd tweet asking the electronics maker why it was advertising on Breitbart News Network, a right-wing website known for scorched-earth populism.
A banner ad promoting the company’s new Linksys mesh router had appeared on the site, even though Breitbart wasn’t among the roughly 200 sites Belkin had preapproved for its ads.
Mr. Hannon called his ad agency, which couldn’t explain the mix-up.
“We still don’t know how that happened,” he said.
Such headaches are becoming all too familiar for marketing executives, as they come to grips with the trade-offs inherent in automated advertising. Known as “programmatic” ad buying, it is now the way the vast majority of digital display ads are sold.
Programmatic advertising allows the buyer to target consumers across thousands of sites, based on their browsing history or shopping habits or demographics. Doing so is more cost-effective than buying more expensive ads on a handful of well-known sites.
But marketers don’t fully control whether their ads will show up in places they would rather avoid: sites featuring pornography, pirated content, fake news, videos supporting terrorists, or outlets whose traffic is artificially generated by computer programs.
The confusion stems from the convoluted infrastructure of the ad-technology world: a maze of agencies, ad networks, exchanges, publisher platforms and vendors. Instead of buying space on websites, brands can buy audiences—categories of people—and their ads are placed on sites those people visit.
The problems arise when those people are on sites where brands don’t wish to appear.
As the issues pile up, marketers are taking action, with the help of companies that independently verify that their ads aren’t…