Almost a quarter of video ad and 11 per cent of display ad views come from bots and fake consumers, with publishers who buy sourced traffic from third parties particularly vulnerable.

The figures come from a study by the Associate of National Advertisers (ANA) and online fraud detection firm White Ops, who analysed 181 campaigns and 5.5bn impressions over the course of 60 days.

Bot fraud originates from malicious sites using phony ad traffic that is passed through both legitimate and “phantom” elements of the digital advertising ecosystem, with fraudsters collecting payments from advertisers for the non-human impressions.

The report estimates that fraudulent traffic will cost advertisers up to $6.3bn (£4.02bn) during 2015. It did identify patterns in bot traffic, indicating that peak traffic occurs between midnight and 7.00 am, when real users are sleeping, and older browsers such as IE6 and 7 showing much higher levels of fraud.

“This study confirmed some prior assumptions and fears, but it is not about sowing distrust or policing ecosystem partners,” said Michael Tiffany, CEO of White Ops. “Its about stopping outright criminal theft.

“Ad fraud is hugely profitable and is one of the major sources of funding for a global underground responsible for a broad spectrum of cybercrime. To protect this cash cow, adversaries are aggressive, smart and adaptable. As such, the results of this study should not be about building better mousetraps, but about driving substantive change in the industry to alter the economics for criminals, and ultimately drive them out of business.”