With the gradual disappearance of user dividends, the mobile advertising industry tends to develop steadily. The industry cheating, which has been criticized by advertisers for a long time, has also reached the deep water of reform. Coping with false traffic and various cheating means is also one of the core functions among the mobile advertising platform. As the “consumer” role in advertising, advertisers hate traffic cheating. In the face with increasingly rampant traffic cheating, how can we “anti cheating” to protect our own benefits?
There are many ways to cheat on Internet traffic, and the most common can be divided into three kinds.
1、Traffic cheating
Advertising exposure, click, conversion may be used to fake amount of cheating.
The downstream network alliance channel makes use of testing machine simulator or platform vulnerability to create user IP out of thin air, complete the behavior of watching and clicking advertisements, and even automatically simulate the behavior of registration and downloading to deceive advertisers.
2、Swipe quantity cheating
This kind of traffic belongs to the real quantity, but the final conversion of purchase or retention is particularly poor. The principle is to use some deceptive or incentive advertising to attract users to click. For example, the use of pornographic material and “you have a red envelope to receive”, which belong deceptive or incentive material.
Brush volume may also be channel advertising on meaningless large site, or advertising pop-up, advertising coverage.
3、Attribution cheating
Due to cheating, what is robbed in essence is the credit of natural flow. Attribution means that the real transformation effect brought by advertisement is regarded as the credit of the advertisement.
Attribution cheating will be included in other traffic, wherever users are attracted to download, all come from this advertisement!
Common anti-cheating methods
These three ways of cheating are the ways that advertisers often encounter. In fact, we can detect cheating and deal with it with some data.
The most common way to identify cheating is to monitor IP anomalies, conversion rate anomalies, traffic deduplication, ad Click to conversion installation time anomalies.
1、IP abnormal
① Statistics of each group (offer_ id, affiliate_ id) conversion rate (CR) of the same IP in each day, conversion rate > 50%, click > 5, marked as suspicious (empirical data).
② Two days before and after the statistics, each group (offer_ id, affiliate_ id), the independent IP of click is taken from the coincidence degree of two days, and the IP is defined_ DUP, which is the number of independent IP overlapped in two days divided by the number of independent IP not overlapped in two days, rate > 0.3 (empirical data), and the number of conversion > 10 (empirical data), marked as suspicious.
③ Count the number of R1 is equal click divided by independent IP in each country, and then count the number of offers in each group(offer_ id and affiliate_ id) R2 is equal click divided by independent IP number in this country, if (R2-R1) / R1 > 1, and click > 1000, conversion > 10, (empirical data) , marked as suspicious Test data.
2、Abnormal conversion rate
Statistics the conversion per hour of each group (offer_ id, affiliate_ id), If the conversion rate is more than 30% (empirical data show that the better conversion rate is 5% – 10%), and data of offer_ ID and affiliate_ ID is greater than 1 / 40 of the total conversion amount in the whole day (temporarily take the empirical data less than 1 / 24 of the hourly average value) *, which is marked as suspicious Text data.
3、The time from ad Click to installation is abnormal
Statistics the time from ad click to installation of each group (offer_ id, affiliate_id), If the number of clicks below x seconds is about XX%, it is marked as suspicious; if the number of clicks above XX seconds is about XX%, it is marked as suspicious.
4、Flow de duplication
Some channel systems will repeatedly send a click after it occurs. If a large number of repeated clicks arrive at the advertiser system, it will affect the advertiser’s evaluation of traffic quality, so it is necessary to de duplicate.
In view of the problem of traffic repetition, using FuseClick system can solve this problem well. You can turn on the filter Junker traffic option. In a certain time interval, repeated clicks will be blocked.
This time interval can be configured as short as 10 seconds and as long as 12 hours of Text data.
Third party cheating detection
In order to further support the on-demand detection of cheating in the digital advertising tracking market, maximize cost savings and reduce your losses in cheating, FuseClick has made docking with several authoritative third-party anti-cheating platforms and systems. Now, you only need to simply set the range of transformation information to be sent to the opposite platform in FuseClick to realize rapid detection.
You can send an email to client@fuseclick.com, and send a copy to the service personnel who connect with you. We will assign special person for you to answer and guide the docking work until you can use it easily.
FuseClick is a third-party PAAS platform focusing on providing global mobile advertising tracking, effect analysis and optimization, and operation management for enterprise customers such as global alliance, mobile advertising agencies, media organizations, and mobile game publishers.
