How does invalid traffic detection & filtering work?

Invalid Traffic (IVT) detection & filtering is done through a combination of different systems, among them:

  • Browser UserAgent filtering
  • IP filtering
  • User filtering

User agent filtering

Search engines, spiders, link validators and other automated traffic sources are filtered based on their User Agent. This traffic is entirely ignored and no ads are served.

IP filtering

We maintain lists of IPs that are known to be a source of non-human traffic. Examples include hosting providers such as Amazon and Azure, as well as residential IPs used by bot & spam networks.

User filtering

We also perform analysis at the user level to detect non-human behavior, based on the email or email hash provided by a publisher. Frequency capping is also involved (although not directly as a IVT filtering system) to prevent abuse from a single user.

 

There is currently no difference between sources of traffic when it comes to IVT detection & filtering. The same logic is applied to email, web, and other container types.