Dirty Tricks in Mobile Advertising Industry
Appsflyer, an international analytical company in the mobile industry, together with partners from Forensiq — experts in identifying fraud in online advertising, announced a September 2016 report on fraudulent mobile traffic transactions. According to their data, fraud costs advertisers at least $1.3 billion annually. Given that the mobile industry is constantly growing, we can assume that these numbers will only increase.
Fraud in the mobile context is malicious traffic manipulation, for gaining profit and cheating advertisers. The most common methods of fraud are bot traffic and the use of incent traffic instead of non-incent.
According to the Appsflyer report, about a third of visits to almost any site on the web are generated by bots — a set of scripts designed to mimic the behavior of a living person in a certain situation. The prevailing amount of fraud traffic is concentrated in the countries with the highest payout per installation or action (USA, China, Germany, United Kingdom, Australia), while in countries with low payouts this type of fraud is not so wide-spread. In fact, most of the bot-holders (people who create and distribute bots) are physically in less-developed countries, remotely creating fraud traffic in developed countries. Historically, most of all bot sources are located in Asian countries. The undisputed leader so far is Vietnam.
The question of substitution of non-incent traffic for incent is more difficult. Some experts do not even qualify this phenomenon as fraud. After all, according to the CPI-model of monetization, the goal is to get the application installed, no matter how it happened. The question is whether such traffic is relevant in reality.
Imagine that you are a game app developer, applying to an advertising agency to increase the number of your product’s active users. You are ready to pay $3 for each new user. Partner offers each user $1 for installing your application. As a result, you get people with zero interest towards your application instead of real active users. Meanwhile, your partner will still receive $2 for every one of these “users”.
Of course, such fraud is simple to track and it is easy to terminate such unscrupulous ‘cooperation’. At the same time, if both kinds of traffic are used, it becomes more difficult to evaluate a partner’s integrity.
So, there’s the problem. How do we deal with it? Basic ways of fighting fraud are:
- Permanent traffic analysis: tracking of duplicate IPs and/or devices, abnormally high conversion rates and other signs of fraud activity. Regular updating of IP “black lists”.
- Comparison of the behavior of new users with the benchmarks of the previous periods. Fraud affects the patterns of user activity.
- Updating the publisher’s “black lists” database and cooperation only with trustworthy partners.
It is important to remember that not every IP duplicate or suspicious action is a fraud, at the same time not every unique visitor is certainly real.
Each advertising agency working with large volumes of data traffic should use an anti-fraud system. Tapgerine uses its own solution to monitor fraud traffic. A comprehensive risk assessment is carried out with each account on a wide range of data analysis: traffic is checked by IP, geolocation, and information on IIN. Using this system, Tapgerine monitors clicks and possible IP spoofing via data centers, updates black lists, and much more.
How to eradicate traffic fraud? Those are hot-topic questions, but the answers are yet to be found. Like any other malicious activity, bots and other types of fraud are constantly evolving. When there’s ten new anti-fraud systems, there’s ten new methods to circumvent them.
Effective methods of dealing with fraud should be aimed both at improving daily software protection and leveling up the business culture. All parties must understand that choosing fair and transparent ways of doing things will bring more profit and normalize the industry in the long run. This understanding is a direct consequence of a person’s high level of culture. Its development is challenging, ambitious and an almost impossible thing to do for only one party. Eradicating fraud requires united efforts of the entire industry, the cooperation of all parties involved and maximum work transparency.