【SpiderAF Case Study】Using SpiderAF as a Third-Party Tool to Strengthen Against Ad Fraud!


Zucks Co., Ltd. Affiliate Business HQ


As one of the largest smartphone-specific CPI ad networks in Japan and a CPA ad network with a large number of projects, Zucks Co provides

performance-based affiliate marketing for posts on blogs and sites. Early on, Zucks Co. developed and used their own fraud detection but now use SpiderAF as a third-party tool. Today we’ll be talking with Mr. Misawa and Mr. Oshima from Zucks Co


Purpose

・While strengthening our own measures against ad fraud, we felt the need to look at this from a third-party point of view

Problem

・Reducing man-hours in our own development

・Necessity for third-party perspective on surveillance and info 

・Visualizing data


Result

・Data that couldn’t be visualized could be seen with SpiderAF

・Using a third-party tool helped boost strengthening operations and detecting unauthorized carriers

・Gained the ability to share detected personalized fraud within the team



How are you going to visualize large measured data



– Please tell us about when you felt the need for fraud detection before introducing SpiderAF



Mr. Misawa

Up until now we’ve used our own fraud detection at Zucks Affiliate. Now when we try to push forward fraud prevention, we have to think about the importance of reducing our own development man-hours and surveillance/info from a third-party’s point of view which is why we thought about introducing SpiderAF.



Mr. Misawa

Then we were thinking about the issue of how to visualize large measured data.



– What kind of visualization were you expecting?



Mr. Misawa

Specifically we were thinking about CTIT (Click to install time). We wanted to visualize the time between a click to installation but we ran into some difficulties at our admin screen when we tried to figure out how to graph this out when dealing with such large quantities of data.



Boost fraud detection from a third-party perspective!


Mr. Misawa

We use our own ZucksAffiliate fraud detection + SpiderAF as a third-party tool to help strengthen operations. With SpiderAF, I can look at data that couldn’t be visualized at our company. Through this I got new information that I hadn’t noticed until now and began using it in our operations. 


For each publisher there is a fraud score listed out; since SpiderAF is linked with our company’s admin screen its easy to compare. In addition, because we can explain this from a unbiased point of view we can push fraud decisions when we’re explaining this to publishers.



Be able to share detected personalized fraud within your team


Mr. Misawa

There were times when I felt that something was suspicious and thought it was fraud but had a hard time sharing within my team since it was personalized. Now since I can visualize the data, it’s been much easier to share this within my team!


– What are some of the most useful and frequently used features on SpiderAF?



Mr. Misawa

I often see the batch saying that my device is old. When I start thinking “Weird. It keeps coming up but nothing happens” I try to find out what device it is and check with our company’s admin screen to confirm. There is a mutual supporting relationship not only with our company, but also with SpiderAF.




Fig : Device batch (Each device has an overseas / old batch so you can make a decision at a glance)



Mr. Oshima

I feel that it is very useful because I can tell by instinct when it comes out in the hour, minute and second in the new report graph where i’ll think “Wow that looks like a bot.” It’s not possible to make a decision based on the first shot if there is no figure, so I use it as one of the general criteria for making a decision.




Fig: CTIT (The time between clicking on an ad to installation of an application. It’s normal that there would be variation depending on the person but it would be abnormal for someone to concentrate on something for say 10 seconds. It shouldn’t be unreasonable to complete a series of tasks in a couple seconds.)




Fig: Clicks per second (By looking at the number of clicks per second, you can easily identify a Bot cycle at a glance.)



– Please tells us about your future business/service prospects working with SpiderAF!


Mr. Misawa


In the future, we hope that automatic systematization will get advanced enough that it will be linked to our admin screens and won’t require manual work. We want to continue strengthening cooperation between our company’s fraud measures × SpiderAF in order to wipe out fraud and provide a safe platform to use.


Within the industry, even a company like Zucks Affiliate – whose been tackling fraud measures from the beginning – rates this third-party measurement tool (market) very highly.




Thank you very much