Ad fraud: Can AI and blockchain help?

With the increasing dynamism of digital media and content platforms, the demands placed on the trade for digital advertising space also increased over the years. Everything had to work faster, more directly and more efficiently. This was no longer possible with the classic “insertion order”. Therefore, a technology for the fully automated trading of advertising was developed – programmatic advertising. This allows the current requirements to be met, but at the same time there are other problems that should not be neglected.

Black box enables ad fraud and endangers brand safety

The supply chain in programmatic advertising is not as linear and clearly structured as it was in traditional commerce. For the data enrichment process for targeting alone, the number of market participants has practically exploded in recent years. It’s no wonder, because there’s a lot of money involved: after initial hesitation, programmatic ad spending is now on the upswing in Germany. According to data from eMarketer, two thirds of display ads with a volume of 1.44 billion euros will already be traded programmatically in 2018. For 2019, the figures will rise to 74 percent and 1.66 billion euros. In the USA, this is already 82.5 percent at a volume of 46 billion US dollars this year.

Being part of the supply chain can therefore pay off. But alongside many serious ad tech providers, there are also some black sheep in the large herd. They are taking advantage of the fact that the programmatic advertising supply chain is largely non-transparent and acts like a black box from the outside. This provides numerous possibilities for ad fraud:

• Domain spoofing: Fake addresses give the impression of belonging to a valuable publisher.
• Bot traffic: Visits are generated by bots, no human actually sees the ads contained.
• Bot networks: Automatic malware acts without the user’s knowledge and creates fake traffic.
• Zero ads: Here, the visibility of the ads is zero because ads are superimposed or the size is reduced to one pixel.
• Ghost sites: Specially designed websites that look perfect for the ad ecosystem but do not achieve organic interaction.

This list could go on and on, but would never be complete. Click scammers are constantly coming up with new methods to deceive not only ad technologies but also ad fraud detection systems. This is why it is very difficult to analyze the current ad fraud situation, as two different figures show: while the Association of National Advertisers (ANA) forecast ad fraud amounting to 6.5 billion US dollars for 2017, the American advertising expert Bob Hoffman arrived at a figure that was around ten times as high.

And what does this have to do with brand safety? Bob Hoffman puts it perfectly in a nutshell: “Anyone not comatose knows that in the bizarre world of ad tech and programmatic buying advertising can show up anywhere. Regardless of the empty promises of agencies and publishers, advertisers cannot control where their advertising appears.”

So here, too, there is a lack of transparency and control. Still.

Can current technologies help?

AI and machine learning are already being used on a huge scale to detect ad fraud. However, this only works to a limited extent, because as soon as a fraud method becomes known and is included in the prevention measures, new methods are developed elsewhere.

Blockchain technology is also repeatedly seen as a solution to ad fraud. This seems to make sense at first, as blockchain promises exactly the things that are still missing in programmatic advertising, such as transparency and security against manipulation. In blockchain advertising, each participant could follow the complete journey of each individual advertising booking and examine it for ad fraud. On the other hand, all instances involved would have to participate completely in blockchain. Whether this will succeed seems doubtful, given the fact that the supply chain is non-transparent, at least in part.

In order to restore credibility, the reputable part of the exploitation chain is stepping up the fight against ad fraud with various approaches. The programmatic platform provider AppNexus, for instance, is offering its customers a refund, if ads are demonstrably played out in an unlawful environment. Other DMEXCO exhibitors, such as Meetrics, want to take a preventive approach to ad fraud.

The bottom line:

Ad fraud and the resulting lack of brand safety is by no means a new problem and will no doubt occupy us for some time to come. Although there are more and more efforts to curb click fraud, the many profiteers make this an uphill battle. Technological hopefuls such as AI and blockchain will hardly change anything in the foreseeable future. Only the move away from easy-to-manipulate KPIs, such as click rate, towards key figures based on real interactions could put advertising back on track in the long term. However, this would require a radical change of thinking from everyone involved. The similar fight against affiliate fraud could act as a similar role model.