This website stores cookies on your device to improve your website experience and provide more personalized services to you on this website and through other media. To find out more about the cookies we use, see our Cookie Policy.If you click "Accept", we will use your information to serve you better with improved and customized services and for marketing purposes. If you click "Decline", we will not track your personal information on the website except to the extent required to honor your preference so that we do not have to cause inconvenience by asking you to make this choice again unless you clear the cookies.Learn more

Can Ad Fraud Be Destroyed Once and For All?

Posted on June 08, 2018
By Matt Kaplan, Content StrategistContent Strategist

All is not rosy in the digital advertising space. Increasingly, fraud is becoming a major concern to advertisers. Ad fraud is expected to cost brand and performance advertisers $19 billion - $51 million a day - in 2018, according to Juniper Research, which is equal to around one out of every $10 of all digital ad spend.

The problem is especially acute with video. The percentage of firms worried about fraud specifically on their video ad networks rose from 33 percent in 2017 to 48 percent this year.

This is far from a new problem, as fraud is nearly as old as time. Fraudsters follow the money, and as in-app advertising spend rises, nefarious players will continue to want to illicitly grab a cut. The questions for everyone in the digital ad space then becomes, what can be done to combat ad fraud, and can it ever be stopped for good?

Know Thy Enemy

This famous idiom from Sun Tzu’s Art of War may be referring to military battles, but it’s just as apt in the battles against ad fraud. The problem is, when it comes to fighting fraud, everyone’s got their own definitions - and these can often be conflicting.

For instance, consider how some would use Mean Time to Install and Retention Rate (daily/monthly usage) metrics to identify fraud. If an ad network sees a sudden spike in installs (low MTTI) but then a low retention rate after the first month, it could very well determine those installs to be fraudulent. But, such a knee-jerk reaction could be premature. After all, that spike in downloads could have been due to a well-timed and effective promotion, but the nature of the app (i.e. a hotel-booking app or car rental app) could mean that it is just not used with great frequency throughout the year.

Everyone in the ecosystem will have different definitions of ad fraud, and often these parameters are self-serving. First and foremost, all supply chain partners need to agree on what actually constitutes ad fraud before it can be combated effectively.

Why Most Current Fraud Fighting Methods Fall Flat

In many ways, ad fraud prevention takes its cues from anti-virus software. First, a threat is identified, sometimes because it has already infected a network and caused damage. Once it’s been identified, an effective tactic to quarantine that threat is developed and then rolled out.

The problem is that this is reactive in nature. As soon as an issue is identified and neutralized, fraudsters have already developed another way, an alternative mechanism, for achieving their nefarious ends. This cycle then just continues ad nauseam, without the good guys making any real headway.

How to be One Step Ahead of Fraud

Instead of being reactive, a proactive approach that anticipates and stops problems before they ever occur is far more ideal. Only such a methodology can capably prevent fraud from happening and causing havoc in the first place.

But, getting to this stage is far easier said than done. Achieving such foresight requires the use of advanced machine learning algorithms able to model and predict all possible fraud scenarios, so that effective anti-fraud mechanisms can be continually developed and implemented in advance. This kind of artificial intelligence requires a significant time and monetary investment, in addition to enough data to make the ML models valuable and useful.

Thus far, AI and ML may be the best bet in effectively fighting ad fraud. It’s possible that a blockchain-based public ledger/database will help keep everyone informed and enlightened in the fight against fraud, but its use in such a capacity is still quite nascent and may not ever be proven totally effective.

Drawing The Line

Fighting fraud also involves taking hard lines and potentially unpopular stances. For ad networks, that may mean cutting ties with previously profitable publisher and advertiser partners until they clean up their acts in relation to fraud. It also means both advertisers and publishers need to be diligent about sharing data for fraud-fighting efforts and cutting ties with known fraudulent parties. Everyone in the ecosystem has a role to play.

It’s also not a one and done investment nor is it cheap, as fighting fraud continuously requires constant vigilance and significant resources. But, through these actions, ad networks can better guarantee ROI along with a more trustworthy and effective digital ad ecosystem.

To learn more about InMobi’s efforts to fight fraud and ensure greater trust, visit inmobi.com/trust.



Back to Top