No one could say that the world’s attention isn’t genuinely hooked by the US election, but even this hot topic isn’t safe from fraud. Recent analysis shows over a third of Twitter accounts following the elections are fake, which means marketers banking on the social media platform to engage an estimated 7.4 million voters could be throwing away their precious budgets on bots.
From ad stacking to ghost sites, digital ad fraud is expected to cost advertisers $7.2 billion this year and research from OpenX and ExchangeWire indicates fraudulent activity within the advertising technology ecosystem is still a major concern for digital marketing professionals. Respondents believe the current rate of fraud is 27%, which is considerably higher than the 5% most are willing to tolerate.
But advertisers shouldn’t be afraid of fraud according to Scott Knoll, CEO and President of Integral Ad Science. Speaking with The Drum at dmexco Knoll explained advertisers should be concerned about fraud because it is affects the entire the global market from London to New York — and its tendency to follow ad dollars impacts emerging technology, established and start-up businesses alike. But it shouldn’t be a barrier to going digital. There are technologies available to minimise the volume of fraudulent impressions advertisers buy and as long as they use the best ad tech tools possible to remain one step ahead of the fraudsters, there’s no need for fear.
There are two approaches to fighting fraud, macro level analysis, which takes a big data approach to identifying large-scale activity, and micro analysis, which examines each impression separately to determine whether it is fraudulent.
Macro fraud analysis
Big data analysis processes vast volumes of impressions from a wide spectrum of channels and platforms to identify irregularities that may signal fraudulent activity. Macro analysis looks at factors such as browsing patterns, page interactions, user data, and inventory source. It may also take into account the geographic distribution of audiences using geolocation data provided by companies such as Digital Element — a method that also has useful applications in the financial sector, providing fintech companies with a means to quickly identity credit card fraud.
Micro fraud analysis
Also known as session analysis, micro analysis takes place in real time to identify suspicious signals during browsing sessions and eliminate specific threats. Signals analysed may include page structures and styling, dynamic browser features, and specific malware identifiers.
While the ad tech ecosystem has plenty of providers offering anti-fraud solutions, advertisers should look for a provider that uses both macro and micro analysis in tandem to detect and prevent ad fraud. They should also make sure their provider takes a proactive approach to prevention, with blocking capabilities that stop fraudulent impressions hitting the server, rather than simply monitoring fraud after the event.
In addition to using the right ad tech tools to avoid buying fraudulent impressions, advertisers can also revise their digital strategy by adapting buying models to make them less susceptible to fraud. Instead of paying for impressions or clicks that are easy for bots to replicate, advertisers should optimise campaigns to real business outcomes such as sales or conversions that are more complex to fake. Mobile advertising research indicates ads bought on a CPM (cost per thousand) basis are ten times more likely to be fraudulent than those bought using CPI (cost per install).
Ad fraud may be a constant source of dramatic headlines and for those affected it can be a media relations headache, but while it is important to be aware of; it is not something to fear. The ad tech industry is continually developing new tools and techniques to minimise the impact of fraud and as long as advertisers make use of these they have nothing to be afraid of.