Mobile MarketingDigital Marketing 

How To Prevent Fraud in Mobile Marketing? 4 Expert Tips

The figures of mobile fraud are staggering! Here are some selected songs: 75% “unexplained” clicks for Apsalar  ; 40% fraudulent clicks according to Trademob  ; 1 Billion dollars of marketing budget wasted in 2015 with the bots, which generate false facilities of apps according to Forensiq …

“We have been experiencing fraud for years, but we are experiencing a steep rise in fraudulent practices and traffic since April 2016,” says Hélène Queriault, Addict Mobile VP Operations. ”  All the advertising agencies we work with are unanimous: the fraud explodes in 2016 . Fraudsters use increasingly sophisticated technological means, and it is becoming more and more technical to identify them.  “

Marketers: How not to get yourself through fraud in the world of mobile marketing?

FRAUD: WHAT YOU, ADVERTISER, SHOULD NOT PAY!

Fraud is generated by publishers (ie owners of mobile applications or mobile websites) that are fraudulent. They claim to display your ad on their app / site, and generate clicks and / or downloads that are paid for by the ad serving your ad.

There are three types of fraud:

  • False traffic  : generated by bots (robots). These are computer farms that simulate the generation of clicks and installs;
  • Diverted traffic  : These are clicks and real installations of your application, but you do not want them. For example, downloads in the Philippines or Mongolia when you want to download them in France. Or incentive downloads when you want quality downloads, downloads realized and counted several times on the same mobile, etc. ;
  • Organic Fraud  : This is your natural traffic that is diverted and charged to you as paid traffic! For example, millions of fake clicks are generated by a virus (click spamming) on ​​millions of advertisements for hundreds of applications such as clicking on the closing cross of an advertisement and the virus counts a click. And if you ever want to download the application X naturally, the virus will count the download as coming from its (false) click.

HOW TO GUARD AGAINST FRAUD?

Upstream, before starting your campaigns:

  1. Pay extra attention to the selection of your partners ! If you buy traffic from qualitative sources, you will be less exposed to the risk of fraud than working with incentive traffic. Nevertheless, all ad networks may be victims of fraudulent publishers, even the most reputable ones like AdMob and GdN of the Google network. Anyone today can download the Google SDK and post advertising in its mobile app / website, without control. It is therefore only a posteriori that we can identify fraudulent publishers. To do well, your selection of partners must not be limited to the level of advertising agencies, but go down to the level of publishers , within each board. At Addict Mobile for example, there area database (DMP) to select trusted publishers.
  1. Use the anti-fraud features of your tracking tool. Some tracking tools can be used to filter fraudulent behavior, for example blocking certain IP addresses and identifying the click spamming.

After starting your campaigns:

  1. Keep track of your visible KPIs: clickthrough rates, download rates, conversion rates within your application (login, session, registration, purchase). As soon as a KPI is out of the average, fraud can be suspected. For example, if the conversion rate within the application is abnormally low, incentive traffic is suspected. If the install / click rate is abnormally high, there will also be suspicion of incentivized traffic or bots. To be effective and detect abnormal behavior, it is important to look at these KPIs at the level of each publisher (within each advertising network)! And beyond the monitoring of KPIs, one must be vigilant on any spike in volume suddenly, 24-48h: any change in fast pace must be watched carefully.
  1. Once an abnormal behavior is identified, analyze the campaign data in detail on the relevant publisher. This data comes from your tracking tool.

We identify false traffic by looking:

  • IP addresses: if the same IP address (or a pattern of similar IP addresses) has generated a large volume of downloads;
  • Device ID: if the same ID device generated multiple downloads;
  • The names of the mobiles (“Claire’s iPhone”, etc.): if the same name is found at a high frequency;
  • Mobile types / brands: if a large volume of downloads is generated on only one type of phone;
  • The delay between the click and the install: if it is too close to zero.

One identifies diverted traffic, that is to say unwanted quality by looking at:

  • Average time of action or use of VPN: to identify traffic from another time zone, from an exotic country;
  • Jailbroken mobiles: to identify traffic not counted by the Appstore;
  • Conversion rates on your mobile website or within your application too low: to identify unwanted incentive traffic.

Related posts