LONDON: More than a third of mobile programmatic traffic is at risk of fraud, a new report has claimed.

According to AppLift, a Berlin-based mobile ad tech business – based on a quantitative study of first-party data from DataLift, its proprietary media buying platform – 34% of mobile traffic is at risk of fraud.

The report, Fighting Mobile Fraud in the Programmatic Era, went on to distinguish two categories of fraud: suspected fraud accounted for 22% of mobile traffic and high-risk fraud for 12%.

Further investigation revealed that campaigns run on a CPM basis are at much greater risk than those operated on a CPC or CPI basis.

So, traffic from campaigns run on a CPC basis was found to be three times less likely to be fraudulent than for CPM campaigns.

And traffic from campaigns run on a CPI basis was ten times less likely to be fraudulent than for CPM campaigns.

The report also found that the relative amount of fraud increases during night-time hours as many anti-fraud tactics are not sophisticated enough to take into account actual levels of app usage.

Tim Koschella, CEO and Co-founder of AppLift, noted that the cost of advertising played a significant role in mobile marketers' budgets.

"The rise in fraudulent activity not only undermines the integrity of the advertising industry, but it is extremely costly for advertisers," he said.

Marketers need to better understand their mobile programmatic data, he argued, and by utilising systems to detect fraudulent patterns prior to bidding they could "salvage a significant portion of their ad spend".

That could run into billions, as eMarketer has estimated that mobile programmatic media buying in the US alone will reach $20.45bn by 2017.

Earlier this year, however, a study from the Interactive Advertising Bureau's (IAB) Mobile Marketing Center of Excellence found that while three quarters of top-level brand marketing executives thought mobile programmatic an important development, just 27% were currently buying this way.

Fraud concerns of the sort highlighted by AppLift could discourage rapid uptake.

Data sourced from AppLift; additional content by Warc staff