Translated with AI from the Russian original.
The updated algorithm can identify users with suspicious activity and bots and block their transitions to advertisers' websites.
The updated Yandex anti-fraud system identifies and blocks some fraudsters even before the ad is shown. Previously, the system filtered "bot clicks" after the transition to the site, but now it analyzes behavioral signals and prevents the display of ads to bots and users with suspicious activity.
If a fraudster remains unrecognized at an early stage, anti-fraud will work when the ad is clicked and show such a user a "captcha". A successfully passed check opens access to the site - but the advertiser does not pay for such clicks and conversions. Fraudulent conversions also do not participate in the training of the advertising system algorithm. Yandex emphasized:
Thus, following the launch, the number of fraudulent conversions decreased by 2 times, and the number of fraudulent applications decreased by 3.5 times. At the same time, advertisers still only pay for real actions of real users.
