Students Combat Click Fraud

6 comments
Source Title:
Getting To The Point: Students' software targets ad fraud
Story Text:

Interesting story about an adaptive click fraud technology developed by three Boston based students. The system is called ClickFacts and sounds worthy of a little attention.

Three college students have decided to lend a hand. Mikhail Ledvich and Mikhail Gurevich met at Boston University, and Gurevich's cousin, Greg Gurevich, joined from Lehigh University to develop a software program called ClickFacts. The program monitors web sites for suspicious traffic and clicking patterns. The three partners received $18,000 this spring from Y Combinator, a Cambridge-based firm that provides funding for startup businesses.

and on the adaptive nature of the software...

The program, unlike other click fraud detectors, teaches itself to recognize new patterns, said Greg Gurevich, a 20-year-old information systems engineering major who developed the algorithm used for the ClickFacts software. ‘‘It adapts to click fraud,'' he said, instead of relying on users to review traffic patterns, a time-consuming process.

Comments

Interesting

It certainly sounds interesting, but recognising patterns feels a little flawed as a concept they can apply to clickfraud - so far as I understand it, there is no single method of clickfraud, but I should think there are a multitude of different human bahaviour patterns when it comes to clicks.

Even more problematic is perhaps finding a way to run the script on third-party sites where the actual clickfraud is taking place...

The program looks for

Quote:
The program looks for suspicious traffic patterns, such as excessive clicks from one computer or clicks that recur every 30 seconds.

If that is the starting point of the technology, I don't see how it could teach itself to perform at an adequate level. Any good click frauder uses a network of computers and ip addresses and will insert semi-random time delays in between clicks and will spend some time on the website in order to look like a real user. Good luck to clickfacts.

First Post

Hi,
Brian and Goodroi Thanks for your posts, we have done extensive research about many types of click fraud, you are both completely right in that there is not one type of click fraud and that several types are very hard to catch, but that’s what we are here for. The article basically oversimplified our system because its not a technological newspaper, but our software is not a simple pattern recognition system. (We have a couple of tricks up our sleeve)

Greg at Clickfacts.com

Testworthy

That really sound interesting to test out.

we're in beta but everyone is welcome to check it out

We're still operating in beta but everyone is welcome to sign up and use our system: http://www.clickfacts.com/1/

The more beta users and feedback we get the better. Of course future perks will be offered to beta users when we go non-beta.

Best,
mikhail at clickfacts.com

Signed up!

I signed up, will be interesting to see.

Kim

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.