Mobile advertising fraud: how to knock it out
With the advent of smartphones, digital mobile advertising has become the new El Dorado for advertisers. Interactivity, accessibility, advanced targeting capabilities… the advantages of mobile are numerous and explain the considerable advertising budgets invested in this medium. But, almost expected and as was the case with desktop advertising, the explosion of this young industry has been accompanied by a very real threat for advertisers: FRAUD. Less known to the general public – because it does not affect end consumers, mobile fraud has a considerable cost for advertisers: $ 1.3 billion in 2015 and this only in the United States, according to the latest study by ‘Ernest & Young. What does this fraud consist of? How can we identify fraudulent actors? Where do they come from? How to counter them?
What is fraud?
In order to better understand what mobile fraud is, let’s quickly go back to the basic principles of digital advertising. An advertiser wants to target smartphone users in order to communicate about their brand, application or site. For this, he will pay a fixed amount to a publisher for carrying out an action on his traffic (impression of an advertisement, click, download an app, registration on a site, etc.) The publisher, who generally offers free content on its application or website, is therefore offered a unique opportunity to monetize its audience and thus build a sustainable business model. So far so good. But now, unscrupulous publishers have decided to “take the stake” by simulating false actions on their traffic. The advertiser, who wanted to acquire qualified users, therefore finds himself with a non-negligible share of unqualified or even non-existent users!
The direct and mathematical consequence is that the advertiser ends up with an actual cost price that is much higher than expected. However, its budgets are not expandable, it will have to exert deflationary pressure on prices. It is then the entire chain of publishers who pay the consequences… fraudsters included!
Detecting fraud, a daily struggle
Today there is a whole panoply of practices to detect fraud. It is an exercise that is both simple and complex, combining rigor and technology. Different common sense principles based on professional experience allow us to obtain very good results and to propose a systematic approach with the objective of eliminating the potential risk of fraud. It is therefore not a question of finding the magic formula that will make it possible to avoid fraud but of combining several approaches in order to make the work of fraudsters more and more complicated until they decide themselves. -even to abandon or attack other weaker prey.
The most classic fraudulent schemes and some approaches to counter them:
The robots take control … No that’s not the title of the last Spielberg. Some hackers have indeed developed robots capable of “watching” advertisements, clicking on them and even installing applications. These techniques are in fact the legacy of a long practice of desktop fraud and today represent the majority of cases of fraud on mobile devices. A solution? The best way to counter these robots is to analyze the campaigns in real time: from which terminals are the advertisements viewed, which IP addresses, which terminal language, which redundancy for the same action on the part of a user, what timing between click and action, etc. So many data and parameters to identify suspicious trends, proxies and robots, and therefore block the source at the origin of fraudulent traffic.
No robots, but exceptional campaign performance for very low quality? The advertiser is probably the victim of an equally common fraud: incentive traffic (ie the user receives a premium for the action). This time, the users are indeed “real”. Only they are “paid” by the publisher to view, click or take action. Results: very interesting conversion rates, but a very low level of engagement or brand loyalty. Here again, an analysis of the traffic (hourly conversion rate, user engagement rate, etc.) will make it possible to identify fairly glaring trends and to block fraudulent sources.
The users are not at all active, yet it is certain that they do not come from incentive traffic. In this case, the advertiser is exposed to an even more devious fraud: the “farms” of installations set up in certain countries. Employees of these “farms” are paid to install the same application thousands of times from the same terminal by simulating different users. How? ‘Or’ What? Through a very simple operation that the tracking systems cannot yet identify: they install and open the application on their terminal, uninstall it, reset their terminal identifier and repeat the operation dozens of times, all at the same time. hiding behind proxies. Laborious but ingenious. It is therefore necessary to be smart and, once again, analyze the data received, in particular the name and the language of the terminal. Indeed, the terminals generally have a name, and send the language of the machine during installation. For example, if we observe for the same campaign targeting the United States, several downloads from terminals with the same username (eg Michael’s iPhone) and having the same language (eg VN for Vietnam), it is This is most certainly a fraudulent download from one of these “farms” of installations located in the country. Or it’s really that your product is a hit with Vietnamese living in the United States!
This is unfortunately only a limited sample of fraudulent actions observed on the mobile. There are even more sophisticated practices that good algorithms will be able to uncover (click spamming, campaign viewability, etc.).
3 techniques to win your fight against fraud
The ecosystem of mobile media publishers is a real jungle into which malicious intermediaries can easily slip. If it is difficult to see clearly and to point out a culprit, there are fortunately tools and methodologies allowing to protect oneself in a very effective way. Advertisers, don’t be paranoid! Mobile marketing – although still in its infancy – is a world of opportunities opening up to you with new targeting possibilities and previously unseen digital transformations. On the contrary, fraudulent activity is inherently short-term and harmful to the entire industry. It is therefore a safe bet that with time and the implementation of good practices, this epidemic will become much less threatening, as was the case for Desktop.
So show the example and be more cunning than fraudsters to foil their attempts:
Multiply the sources of traffic (RTB, social, ad-networks, etc.), formats (display, video, rich media, etc.) and approaches (retargeting, etc.)
Track and analyze user engagement as well as possible.
Choose the right partner to help you and guide you in media buying.
Any source of traffic includes more or less qualified users, the key is knowing how to sort and remove pollution from fraudsters!
Stephane Pitoun is co-founder and CEO ofAdxperience, a mobile marketing platform that has developed its proprietary programmatic technology allowing advertisers (digital agencies, applications, mobile games, etc.) to connect their campaigns to different sources of international mobile traffic. Stéphane Pitoun has more than 15 years of experience in the development of online media strategies in France and abroad.
Photo credit: Fotolia, royalty-free stock images, vectors and videos
Read also: 3 fundamentals that we do not tell the advertiser in the advertising market