Google is making system updates to fight invalid traffic and suspicious activity on its ad networks, the company announced Wednesday. It will also give AdSense and AdMob publishers more information when ad serving is restricted as a result of these new measures.
New risk prediction models. Google said this launch incorporates new machine learning models to predict high-risk ad traffic and block ad requests before they serve. The aim is to “identify potentially invalid traffic or
“These defenses allow us to limit ad serving as needed to further protect our advertisers and
Site verification paying off. This builds on Google’s existing detection and filtering systems as well as its move last year to require new sites to go through a process to verify domain ownership or authorization to modify a site’s content before being approved to serve ads. Prior to that change, site owners could simply reuse (and misuse) ad code from another site without adding the new sites to their AdSense accounts. Google said Wednesday it now blocks more than 120 million ad requests with this feature.
Publisher notifications. Google says most publishers won’t notice the change, but those who are impacted will be notified about ad traffic restrictions in the AdSense or AdMob Policy Center. In the UI, they’ll see that the number of ads they can show has been limited and recommendations to address the restriction. “This will allow them to understand why they may be experiencing reduced ad serving, and what steps they can take to resolve any issues and continue partnering with us,” said Ferrate.
Publishers that notice declines in ad traffic should check the Policy Center for any notifications related to ad restrictions.
Why we should care. For advertisers, the system updates should provide additional protections against paying to show your ads to bots rather than real people. Advertisers won’t have visibility into which domains are impacted since the blocking happens before ads can serve. Google also isn’t disclosing any specifics about what its systems are designed to detect in order to keep bad actors from potentially reverse engineering its efforts.