Home > General > Affiliate Marketing Forum

ClickFraud.co Beta Launch & Accepting Beta Tester Applications (12)

09-26-2015 08:07 PM #1 juntao65 (Member)
ClickFraud.co Beta Launch & Accepting Beta Tester Applications

We all know click fraud is a big problem. It is estimated that 30% of all web traffics are fake. Lower quality sites that percentage could be much higher.

I was fed up with paying so much for bot traffic. It is obvious some traffic sources were very low quality, but it was hard to pinpoint exactly where the bots are coming from because a good mixture of converting traffic still comes from the same traffic source. I decided it's big enough problem to tackle it head on. Today I am launching a new product designed to track and analyze bot traffic that works with Voluum and any of your existing tracker. It will work with cloaker too. The platform is in active development, and with your feedback it will improve constantly. This is just the beginning, the plan is to build a company out of this idea of defeating click fraud to help internet marketers to make more money. Message me if you are interested in joining the beta testing program.

http://clickfraud.co

Feel free to comment, pm me or email info@clickfraud.co directly.

09-26-2015 08:29 PM #2 cmdeal (Veteran Member)

Perhaps you can describe a little more on your website what exactly you are doing and how your solution actually addresses the problem. Some screenshots and real testimonials would definitely help.

I don't think you need to convince anyone that traffic fraud is a problem.

Rather you need to convince people that what you have is the solution.

If I am being honest, there is really very little substance on your website there to convince me at least that I should give this a shot.

09-26-2015 08:37 PM #3 juntao65 (Member)

Good tips. I post some screen shots and talk a little more about how it works on our website.

The gist of it is that we have an algorithm that analyze the behaviour of each visit and mark it as bot or not bot. There are some behaviours that only bot does and humans won't do. The exact method is proprietary so I cannot disclose as it is the bread and butter of this software. As data grows we will later on also implement supervised machine learning algorithm to further improve the detection capability.

09-26-2015 08:43 PM #4 cz_voyager (Member)

Does it work with both desktop and mobile traffic?

Does is slow down loading?

09-26-2015 08:49 PM #5 juntao65 (Member)

Quote Originally Posted by cz_voyager View Post
Does it work with both desktop and mobile traffic?

Does is slow down loading?
Works for both desktop and mobile.

My testing shows that it adds about 50 milliseconds to the redirect chains, which to most people is fairly negligible.

09-27-2015 03:30 AM #6 zeno (Administrator)

I think there should be a bit more technical information, e.g. how the system is actually implemented - is it a super simple javascript embed, is it a redirect system, do we integrate it with a tracker, etc.

How does this software automate and improve the bot detection process?

The thing is, in the context of improving ROI, I will look at my tracker data and use that to kill off poorly performing segments of traffic. That data is the most important because I can see what is actually generating money.

Does this system make it easier to find those segments by providing a clear, consolidated report? Can it automatically implement rules for a campaign to redirect zero-revenue-generating bot traffic to a black hole?

Another thing that would be useful is providing a traffic-source wide (across clients) anonymous report to traffic sources about what traffic is being detected as bots. A community service.

09-27-2015 07:46 PM #7 juntao65 (Member)

Quote Originally Posted by zeno View Post
I think there should be a bit more technical information, e.g. how the system is actually implemented - is it a super simple javascript embed, is it a redirect system, do we integrate it with a tracker, etc.

How does this software automate and improve the bot detection process?

The thing is, in the context of improving ROI, I will look at my tracker data and use that to kill off poorly performing segments of traffic. That data is the most important because I can see what is actually generating money.

Does this system make it easier to find those segments by providing a clear, consolidated report? Can it automatically implement rules for a campaign to redirect zero-revenue-generating bot traffic to a black hole?

Another thing that would be useful is providing a traffic-source wide (across clients) anonymous report to traffic sources about what traffic is being detected as bots. A community service.
how the system is actually implemented - is it a super simple javascript embed, is it a redirect system, do we integrate it with a tracker, etc.
It has both redirect + javascript embed. You just set the landing page on your tracker the landing page you have setup on ClickFraud. It takes about 3 minutes to setup.

How does this software automate and improve the bot detection process?
We analyze the behavior of each page view and generate a report for you search/sort/export to see which placement is a bot. You can compare against any variable that is passed in through the URL.

The thing is, in the context of improving ROI, I will look at my tracker data and use that to kill off poorly performing segments of traffic. That data is the most important because I can see what is actually generating money.
You still have all the functionality of your regular tracker to see the ROI of each campaign. What ClickFraud does it gives you another dimension of data to see which placement has more bot traffic, which then you can optimize it by blocking say 1 particular site, or say one type of browser signature. For example you might have a profitable campaign already, but it is likely that you are still paying 20% premium for some bot clicks that you can start black listing out to further improve your ROI.


Does this system make it easier to find those segments by providing a clear, consolidated report? Can it automatically implement rules for a campaign to redirect zero-revenue-generating bot traffic to a black hole?
Yes you can analyze against any variable thats passed in through the traffic source, generate blacklist/whitelist and import it into your traffic source.


Another thing that would be useful is providing a traffic-source wide (across clients) anonymous report to traffic sources about what traffic is being detected as bots. A community service.
I thought about implementing this as an opt in feature. Some affiliates might not want to share the blacklist/whitelist they build up over time by running traffic through ClickFraud, but if some are willing to share then we can definitely offer this as a feature.

10-01-2015 11:08 PM #8 revolution (Member)

Looks Interesting...feel free to reach out to me

10-03-2015 04:13 PM #9 azureus (Member)

Hi juntao65,

can you maybe share - a small part obviously - of your shit list, i.e. a list of fraudulent publishers or traffic sources on some platform?

I myself keep my blacklists for pretty much any large traffic source I use (at the moment I have 15+ blacklists).
I'm able to detect most of the fake traffic but it is manual, takes a lot of time and I usually get to it a week or two after I've received the fake traffic. So your tool would be helpful.

It would be nice to compare our results to check if your tool really works.

So I suggest you write some networks on which you successfully used this and then we can exchange our blacklists to see if your system has the same results. You can actually do it publicly as zeno suggested.
You do not need to share your method/algorithm and noone will ask you to do so but if it's legit, you should be able to share a sample of your results.

Also to be honest, it is much more difficult to do this for mobile so I'm very sceptical when you say "Works for both desktop and mobile. " as if it was no big deal. This sounds more like you have no idea about how very different click fraud on mobile is.

10-03-2015 07:41 PM #10 leonidas32 (Member)

Hey Junta, if possible, can you present some data from your own internal testing.

That would be awesome.

10-06-2015 03:19 AM #11 juntao65 (Member)

Quote Originally Posted by revolution View Post
Looks Interesting...feel free to reach out to me
PM'ed you

Quote Originally Posted by azureus View Post
Hi juntao65,

can you maybe share - a small part obviously - of your shit list, i.e. a list of fraudulent publishers or traffic sources on some platform?

I myself keep my blacklists for pretty much any large traffic source I use (at the moment I have 15+ blacklists).
I'm able to detect most of the fake traffic but it is manual, takes a lot of time and I usually get to it a week or two after I've received the fake traffic. So your tool would be helpful.

It would be nice to compare our results to check if your tool really works.

So I suggest you write some networks on which you successfully used this and then we can exchange our blacklists to see if your system has the same results. You can actually do it publicly as zeno suggested.
You do not need to share your method/algorithm and noone will ask you to do so but if it's legit, you should be able to share a sample of your results.

Also to be honest, it is much more difficult to do this for mobile so I'm very sceptical when you say "Works for both desktop and mobile. " as if it was no big deal. This sounds more like you have no idea about how very different click fraud on mobile is.
The thing is many traffic sources have mix amount of fraud traffic. it also varies depending on the type of targeting. for example some bots might be only pretend to certain type of "device" on a placement. With ClickFraud you can find out which targeting the bots are and remove it from your targeting. So the value isn't to make a list of site to block, but a list of more granular targeting rules to filter out the bots.

From web traffic perspective, there is nothing fundamentally different from mobile and web traffic. A client send a GET request, and a server responds with a 200, then client render the page.


Quote Originally Posted by leonidas32 View Post
Hey Junta, if possible, can you present some data from your own internal testing.

That would be awesome.
See my above post.

10-11-2015 08:00 PM #12 azureus (Member)

Lol your answer sounds like a joke. I'm still waiting for a legit and open answer from you. You avoid the question really hard.
"traffic sources have mix amount of fraud traffic". May be true. However, if your tool really works, you are able to detect sources which send e.g. 75%+ or 90%+ bot traffic. Why not share them if your tool really works?

Also your answer about mobile and web fraud traffic differences is not relevant. You have no idea how these people operate.

Home > General > Affiliate Marketing Forum