About Invalid Traffic

Invalid traffic refers to traffic that does not bring any value to the advertisers. It comprises non-human traffic or NHT (i.e., robotic), humans with no intention to convert, and malicious or intentional activities (i.e., fraudulent traffic) trying to defraud advertisers of their marketing budget.

Different types of Invalid Traffic 

As per Trustworthy Accountability Group's (TAG) IVT Taxonomy following the Media Ratings Council's (MRC) Guidelines, the types of invalid traffic are: 

• General Invalid Traffic (GIVT): Traffic detected using routine filtration methods such as lists or with other standardized parameter checks.  

• Sophisticated Invalid Traffic (SIVT): Invalid traffic that requires advanced analytics, multi-point corroboration/coordination, or significant human intervention to analyze and identify.  

How does Criteo filter out invalid traffic?  

To provide a trusted and safe advertising environment, we use a combination of machine learning and in-depth human analysis to detect, prevent and discard invalid bid requests, impressions and clicks.  

Our system uses sophisticated filters to distinguish between human and invalid traffic including: 

• Industry-compliant blocklists (TAG Data Center IP List, IAB/ABC Spiders and Bots List) 

• Our blocklists of suspicious IPs, unsafe publisher domains and app_ids 

• Rules-based filtration (Activity-based rules measure metrics and calculate ratios to identify bad actors) 

• Machine-learning models to detect advanced patterns 

We collaborate with Pixalate, a Media Rating Council (MRC) accredited service, to enhance our ability to detect and filter invalid traffic including blocklists of suspicious IPs, high-risk mobile apps and more, both before and after bids. Additionally, we are TAG certified Against Fraud (CAF) and hold Brand Safety Certification (BSC).