The U.S. Presidential Election has, in many ways, been digital. Spend on digital ads in the race reached $2.9 billion in 2020. This was up sharply from $0.4 billion four years ago, marking the continuing prominence of digital political campaigning since President Obama's campaign manager, David Plouffe heralded the channel as a deciding factor in the election 12 years ago.
However, an increasing challenge for this online ad spend has been ad fraud. In a new study, in association with the University of Baltimore, we see that marketers will $35 billion to digital ad fraud in 2020.
Ad fraud is the practice of fraudulently representing online advertising impressions, clicks, conversion or data events in order to generate revenue. In the case of the political campaigns, often money is spent reaching bots rather than voters.
How ad fraud is getting more sophisticated
The impact of bots and ad fraud on digital ad spend has grown as ad fraudsters have become more sophisticated. For the majority of fraudsters, the automation tools used to commit fraud are evolving without them having to do anything about it – fraudsters just have to hide and rewrite certain elements in order to evade more and more tests. In addition, one in five click fraud cases involve VPNs or proxies – disguising the fact that the user clicking on a campaign ad is actually located in countries other than the U.S. – and therefore almost certainly not eligible to vote.
Bot-makers create millions of headless browsers, that can simulate all human-like actions on ads and landing pages, such as mouse movement, page scrolling, and clicks, to load webpages and cause ad impressions, that appear entirely human. Malicious SDKs for advanced and AI-powered click injection are sold in the Dark Web on sale to the public for a fairly low price to perpetrate ad fraud, offering the opportunity in the words of the suppliers to "emulate ad clicks and hijack clicks including Google, and Facebook and organic clicks."
In relation to our findings around digital campaigns, we expect that ad fraud in digital ad spend accounted for $377 million in wasted budget. Some of the findings include that one in five click fraud cases involve VPNs or proxies – disguising the fact that the user clicking on a campaign ad is actually located in countries other than the U.S. – and therefore almost certainly not eligible to vote.
How digital campaigns suffered from ad fraud
There are a number of reasons that political campaigning runs into ad fraud problems. Firstly, the need to spend money and fast diminishes accountability. Zach Edwards, who worked on the online digital campaigns for President Obama and Mayor Bloomberg, says: "There is no ad fraud department in most of these buying teams that can be spending "$30,000 to 50,000 a day. In a business ad buy you have an ROI, you have people checking it for months afterwards, and if it didn't pan out you may ask for a refund, whereas in political buying the only day that matters is election day. So, you are spending money as fast as you possibly can. Because the political campaigns are so short, there is no accountability for ad fraud. But because the campaigns dissolve, no one ever checks that, and no-one ever asks for refunds."
Secondly, most spend online for political campaigns is at the end of the U.S. campaign. However, limits set by Google and Facebook, the two safer destinations for ad cash, has meant other more junk ad-fraud prone traffic received late splurges of cash. This type of traffic over inflates its numbers through bot traffic. In the words of Cam Cameron, digital political consultant at Strategic Partners & Media: "There is definitely a too good to be true feeling when I am being pitched by new partners, new websites where they are coming at me telling me they are getting so much traffic - a local news blog will come at you and I will do the math on their area and it does not make sense. I have a lot of questions to ask them, starting with, 'How can you prove you do not have bots and that is not a large proportion of your views or clicks?'"
Thirdly, OTT streaming where campaigns have spent millions of ad dollars, is particularly vulnerable to ad fraud. Fraud rates on OTT inventory stand at 17%, according to almost all estimates. Several OTT ad fraud cases has been discovered in 2020.
Finally, competitor clicks (one party clicking on another ads) are also likely. This involves opposition campaigns clicking on ads of their competitors to drain ad spend. With pay per click on political keywords costing up to $10 per keyword, this sneaky ad fraud prevents ads from being seen by the intended electorate. This is despite such click fraud being illegal violating the federal Computer Fraud and Abuse Act (CFAA). Under this law, you can go to prison for up to 10 years. Such competitor click fraud has seen a rapid increase during COVID-19 particularly in white collar sectors such as real estate and law firms, under economic pressure. Click fraud is perpetrated to deplete advertisers' budget in an attempt to remove the ad from the search engine results, and damages campaign data.
Bots don't vote, however large sums of campaign spending in 2020 has been spent in reaching them, rather than real voters. In reality, few campaigns think about carrying out post-mortems in digital spend after a campaign. However, seeking answers on whether campaigns reached voters, or bots, represents a crucial analysis to ensure greater efficiency for future political races.