Mobile & Fraud – Scourge of Modern Times

By Lara Cocco. Published Feb 15, 2019

Mobile Ad Fraud has been in the news recently, but what really is it? 

Who are the victims? And why is It important in the world of advertising?

The investment in Mobile Ads has grown exponentially in the past years, in such a big industry, it is expected that the frauds grow too. According to Forbes, in 2018, the number of mobile ad frauds increased twice as much when compared to 2017. This situation costs billions to advertisers.

To explain what Mobile Ad Fraud means in a practical way, we can use Instagram as an example. If you paid Instagram to promote your post and you start gaining a big number of foreign followers that have only a few pictures on their profiles, you will think that this doesn´t look right, since it looks like these accounts are fake. Those accounts won´t benefit you or your business. You want quality followers that will have a real impact on your brand.

The difficulty with getting good results that are worth your money is that digital advertising is constantly changing, and organizations are struggling to respond in an effective way, especially because when the industry finds a way to neutralize or prevent a form of fraud, the fraudsters come with new and more sophisticated tricks.

The problem with Mobile Advertising Fraud has been a huge thing in the industry of advertising. According to Forrester Consulting, 69% of 250 marketers that spend at least $1 million on digital advertising per month, had at least 20% of their investment wasted on fraud in mobile web. Also, the problem is only getting worse, a report developed by the digital market research firm Juniper, predicted that the waste could be expected to rise in the following years, reaching $44 billion by 2022.

But how this situation can be solved? Despite the negative information, there are solutions to the problem. Artificial Intelligence, popularly known as AI, has been used as a way to detect fraud. Intelligent algorithms can operationalize data in a more effective way, spot abnormalities, and set alerts when needed.

More and more companies rely on AI as a form to detect mobile ad fraud, but still, there is no tool capable of stopping the problem in real-time. For our App-download and activation campaigns is even more tricky, since there are more ways to get fraudulent traffic into app download campaigns such as click injection, install farms, fake device metadata, etc.

We study the different types of user flows that happen in our campaigns, our system learns every second how a real human interacts with the app. How they get there (by checking their unique phone id and which impression id they come from for example), then with all the in-app postbacks of the campaign we can identify when the user flow is not running as it should be (e.g. when appears an install and uninstall in a matter of seconds) and block the traffic, learn from it and don’t let it pass our filter again.

Think about it as an intelligent barrier, at first, we let our campaign run and let the users end all parts of the process (install, register, subscription). It develops different patterns of different users, patterns we analyze and when we find some unusual behaviour, we isolate it and mark it as fraud so the next time when we identify the first step of that bad pattern elsewhere, we don’t let it in, blocking it as fraud.

As always, feel free to like, comment and share your doubts, questions or thoughts.

YD Cheers