← All terms·Cross-media measurement
Reference library · Cross-media measurement
Ad Fraud Rate
Ad Fraud Rate is the share of delivered impressions (or spend) attributable to fraudulent inventory or traffic rather than genuine human exposure — bots, spoofed domains and apps, ad stacking, or falsified SDK signals. It is measured post-campaign by a verification vendor and used to size wasted spend and request make-goods or credits from the seller.
Formula
Worked example
A display campaign in Poland delivers 4,200,000 impressions at a €3.50 CPM, for total spend of €14,700. The verification vendor's post-bid report flags 168,000 impressions as sophisticated invalid traffic (spoofed app IDs and non-human patterns). Ad Fraud Rate = 168,000 / 4,200,000 x 100 = 4%, meaning €588 of the €14,700 spend bought no real exposure and should be reclaimed as a credit.
How it is used
Planners pull fraud rate from third-party verification tags (IAS, DoubleVerify, Adloox) rather than the platform's own reporting, and write a maximum acceptable rate — often 1% for premium inventory, higher for open exchange — into the insertion order so breaches trigger a credit automatically. The common mistake is reading the vendor's blended invalid traffic rate as the fraud rate: that figure usually includes GIVT (known bots, crawlers, data center traffic) alongside SIVT (actual fraud), which inflates the number and misstates what was genuinely fraudulent versus simply non-human but declared.
The common mistake
Use only the SIVT (sophisticated invalid traffic) rate as ad fraud, not the blended IVT rate, since GIVT is routine bot traffic and not fraud.