Google took down 2.3bn ads across its network last year, nearly one billion fewer “bad ads” it had to remove in 2017.
The tech giant also terminated nearly one million bad advertiser accounts, about twice as many it axed in 2017.
Google introduced 31 ad policies in 2018 in response to new scams. Some policies include banning ads from for-profit bail bond providers and higher scrutiny on ticket resellers.
Google removed almost 207,000 ads for ticket resellers, over 531,000 ads for bail bonds and about 59m phishing ads. It’s launching a new policy manager tool next month to help advertisers create compliant adverts.
Google also kicked around 734,000 publishers and app developers off its ad network and completely removed ads from almost 1.5m apps. The company said it was able to “take more granular action” by cutting ads from nearly 28m pages that violated its publisher policies.
According to Google, it also removed ads from around 1.2m pages for violating policies on “misrepresentative, hateful or other low-quality content” in an effort to fight against fake news and misinformation.
To improve its machine learning technology, Google launched 330 detection classifiers to determine an ad’s “badness” directly on a website, around three times as many classifiers it introduced in 2017.
It also worked with White Ops and the FBI to take down an ad fraud ring that generated over 3bn bids daily at its peak.
Google also introduced new political advertising policies in the US ahead of the 2018 midterm elections. It verified almost 143,000 election ads in the US and introduced a political advertising transparency report. Google will launch similar tools ahead of 2019 elections in India and the EU.