Increasingly, advanced bots seek to replicate human behavior patterns.
Ad fraud is a major problem in digital advertising that results in billions of dollars in annual losses to marketers.
While marketers are winning the battle against “basic bots,” they’re losing against more sophisticated bots, according to a new study and report from White Ops and the Association of National Advertisers (ANA).
The study collected data on the ad buys of 49 advertisers between August 1 and September 30 last year. ANA concluded that on average, bot impression fraud had cost the study participants roughly $10 million each. In total, the study’s authors estimate that bot fraud will generate $7.2 billion in losses for marketers in 2016 on a global basis.
Source: White Ops-ANA (2015)
The study found that sophisticated bots were targeting higher-value impressions, including video. For example, there was 39 percent higher bot fraud when CPMs were over $10. The report also concludes that “bots are successfully fooling viewability measures” by largely mimicking human behavior patterns.
A variety of bot “tactics” were used to defeat bot-detection systems now in place:
- Bots exploit users’ cookies to appear as humans in general detection and prevention systems.
- Bots spoofed viewability, showing nearly the same viewable rates as humans. Bots fooled list-based prevention technologies in programmatic buys.
- Desktop bots impersonated mobile devices to consume mobile media.
Programmatic ad buys saw higher fraud rates than direct media buys — in some cases, much higher:
- Programmatic display ads had 14 percent more bots than the study average.
- Programmatic video ads had 73 percent more bots than the study average.
- Direct video ads, where measurable, were 59 percent less likely to have bots than the study average.
- Direct display ads were 14 percent less likely to have bots than the study average.
The majority of sophisticated bot traffic was coming from residential computers, followed by corporate IP addresses, according to the ANA findings. The graphic above shows the distribution of bot traffic across the US. The graphic below indicates the IP category sources of bot traffic.
Where Sophisticated Bot Fraud Is Coming From
Source: White Ops-ANA (2015)
A relatively small percentage of residences where computers are infected with malware are generating most of the bot traffic. Roughly “80 percent of successful bot traffic came from the 2 percent of households with the freshest malware infection” the report explains.
To better combat increasingly sophisticated bot fraud, the report offers a list of detailed recommendations that combine technology with anti-fraud policies and guidelines. Yet it’s clear that even after implementing these recommendations, bot fraud will remain a high-stakes cat-and-mouse game in the industry.
(Some images used under license from Shutterstock.com.)