Oracle says it has exposed the largest ever connected tv fraud operation which tricked advertisers into paying for advertisements that were never delivered to as many as 28.8 million US households.

Dubbed “StreamScam” by Oracle, the fraud operation spoofed more than 28.8 million U.S. valid household IP addresses, including approximately 3,600 apps and 3,400 unique CTV device models.

Update: Oracle estimates the fraud operation has cost advertisers at least $14.5 million.

Connected TV is a relatively new marketing channel and measurement technologies are not widely adopted. Analysis has shown demand often outstrips supply, creating ain ideal environment for fraudsters.

Oracle says the StreamScam perpetrators used forged household IP addresses, app IDs, and device IDs and exploited vulnerabilities in Server Side Ad Insertion (SSAI) – CTV ad serving technology used to improve viewing experiences – to send false impressions that an ad had been served.

Oracle used its advertising analytics platform, Oracle Moat, to count the number of ad impressions that are inserted into video streams by SSAI servers as well as the number of ad impressions that actually play on end-user devices. 

The company says it was able to identify the false impressions sent to Moat, eventually uncovering the operation.

Follow the money

Connected TV (CTV) has been considered a bright spot for digital marketers in 2020 when adspend was pulled back from almost all other channels amid COVID-19 uncertainty.

Marketers are estimated to have spent $8.11 billion on CTV ads in 2020, and are forecast to spend $11.36 billion next year, according to eMarketer which identifies measurement and ad fraud among the biggest challenges of the channel.

“Where advertising dollars go, criminals will follow, and rapidly-growing channels like CTV are presenting new opportunities for ad fraud and theft,” said Mark Kopera, head of product for Oracle Moat. 

Research by DoubleVerify released in August showed CTV as the only device with increasing rates of fraud, including a 161 per cent jump in fraudulent CTV traffic rates in Q1 2020.

Previous post

Twitter goes to the cloud with AWS

Next post

News code still ‘unworkable’ despite concessions, says Google ANZ boss