Why is Device Ranking Not a Good Idea?

Why is Device Ranking Not a Good Idea?

As advertising on mobile plays an integral role in online marketing now, fraudsters have started targeting mobile ads instead of the ads on the web. One of the most popular ways to prevent such frauds is by ‘device ranking’. However, preventing such frauds with device ranking at least is not something that is easy or readily effective. Let us have a look at why device ranking is not such a good idea.

What is Device Ranking?

Device ranking is a system which creates a blacklist of devices which were somehow involved in such frauds in the past. A fraud specialist on MobileMarketingMagazine has defined device ranking as a system which blacklists devices on the basis of the metadata and some events which have tracked from the device in the past.

These systems take their cues from the profiling systems that were used for the desktop web. The reason these systems worked on the desktop web was because fraudsters used a specific device for frauds, like credit card frauds. As a specific device was used, it was much easier to uniquely identify the device in some time.

However, the user acquisition fraud mobile is a completely different ball game. In such frauds, a specific device doesn’t play any role, making such systems highly ineffective. Some of the common problems with device ranking for mobile devices include-

1. The system often fails to prevent simplest of frauds

http://www.glispa.com/wp-content/uploads/mobilefraud.jpg

App developers test the apps by running them in a simulator where the app is tested in a mobile environment that is simulated. However, this is done on desktop computers.

Such simulation software can also be used for fake clicks, post-install conversions, and clicks. This is one of the most common types of mobile fraud- simply click any ad in the simulator, install the app to which the ad takes you in the simulator and cash the payout. Moreover, this process can be almost fully automated can be operated from a datacenter or also on the cloud.

But the kicker is, every time the simulator is restarted, a new device will be available. This means that the device will have a new device ID which is generated randomly and can even have a device type which is assigned randomly. While it can be a Japanese-language iPhone 6 now, it can be a Chinese-language Sony phone the next time.

While you can catch simulated devices by simply looking at them, this can only be done once as the next time, the simulator will generate a completely new profile which won’t match with the previous profile.

2. Device ranking often rejects genuine organic conversions

 As per a study by Adjust, user acquisition frauds on mobile occur in two different ways- the simulated way mentioned above and in other cases, actual devices are hijacked to fake the engagement with the ads which were not even displayed. This second method is known as ‘click spam’ and in a bid to prevent such frauds, the device ranking system often rejects organic conversions.

The hijacking for a slick spam is only done for a short span of time which is adequate to receive credit for the conversion. However, this doesn’t mean that the device itself is illegitimate or the behavior of the user makes him/her a blacklist contender.

If you continue blacklisting the users based on such hijacks, you’ll end up rejecting the channels which are actually bringing organic conversions. As the device itself has no role in the fraudulent activity, rejecting the devices is no solution to the problem. Most probably, you’ll end up undermining legitimate data which can be used by marketers for making decisions.

3. Device ranking intrudes the privacy of users

While analytics have been around for a very long time, they do not conflict with the privacy of the users. However, for the device ranking to work, every single activity of the user’s needs to tracked extensively.

With device rankings, there is no way to collect the user data in a way that respects the privacy of the users. Every single thing that a user does in any of the apps is used to create a user profile which is then used for filtering fraudulent activities.

While a lot of people do know that their activities are analyzed, they are generally unaware of the extent up to which their activities are tracked with device ranking systems. If they do get to know this, this will definitely be a fact that they wouldn’t be able to digest.

Conclusion

Device ranking on mobile is completely different from the profiling of desktop web. In this, individual devices never play any role in the frauds, making it useless to blacklist the individual devices for preventing the frauds.

Marketers should try to understand how the fraudsters actually work while relying on other fraud preventing methods which actually work, like filtering IPs that are anonymous, advanced analyses of the device type and distribution modeling. While blacklisting specific devices is not useful, the statistics behind campaigns or specific publishers can definitely stand out.

In short, user acquisition fraud has nothing to do with the devices and the devices cannot be used to track the fraudsters. Using the ranking system to prevent such frauds means completely missing the actual target.