US invalid traffic and MFA rates lead global benchmarks

The findings prove that brand safety and fraud prevention are decoupled; a campaign can be perfectly 'safe' while remaining highly inefficient. For the streaming and digital video ecosystem, this gap highlights a vulnerability in automated buying paths where high-spend markets like the U.S. become primary targets for bot networks. As AI-generated 'slop' content proliferates, the reliance on content adjacency alone is insufficient to protect margins. Watch for a shift toward 'quality-first' programmatic curation, such as Private Marketplaces (PMPs), which recently saw a 23-point jump in spend share as buyers flee the open exchange's higher fraud risks.

The rise in low-quality inventory aligns with broader industry data indicating a persistent transparency gap in the programmatic supply chain. Per the ANA’s Q1 2026 Programmatic Transparency Benchmark released in May 2026, the gap between high- and low-performing advertisers reached a record 21.9 percentage points. Top-tier marketers now convert 54% of their spend into 'qualified' impressions—those that are viewable, fraud-free, and non-MFA—while lower-performing peers convert just 32%. This divergence is driven largely by the aggressive adoption of log-level data analysis and concentrated supply paths to filter out MFA sites, which the ANA identified as a growing subtype of 'AI slop.' External auditing remains critical as fraud tactics evolve beyond simple bot scripts. Per Juniper Research in March 2026, global ad fraud losses are projected to exceed $100 billion this year, fueled by 'Agentic AI' that mimics human browsing patterns like scrolling and hesitation. In the CTV space specifically, fraud is growing at a faster rate than display, with losses estimated at $3.8 billion due to sophisticated device spoofing. Despite these headwinds, industry bodies like the ANA report that MFA exposure among highly managed campaigns has stayed below 1.1%, suggesting that while the open market is riskier, active curation is successfully shielding premium buyers. Furthermore, the surge in MFA inventory is increasingly linked to automated 'Inverse Optimization.' According to reporting from Spider AF in early 2026, AI-driven bidding algorithms often misinterpret high-volume, low-cost invalid traffic as high-performing, inadvertently funneled more budget into fraudulent placements. This has led to a 1,409% year-over-year increase in the number of identified MFA domains. To combat this, platforms are shifting toward Signal Normalization to ensure that machine learning models are trained only on valid conversion data rather than bot-generated signals.