Which methods do Google (and other PPC companies like Adbrite and CommissionJunction ) use to prevent click fraud ??
Thank you![]()
Which methods do Google (and other PPC companies like Adbrite and CommissionJunction ) use to prevent click fraud ??
Thank you![]()

how do we know?
excuse me... I forgot to say "in your opinion"![]()

Algorithms and shit that people smarter then you came up with.
any smarter answer ?
They track information that is sent by the web browsers very closely such as user agents, referers etc..., then also use cookies to see where you have previously been in they're system. They might also use javascript and ajax to report information to their servers. Tons of ways they can tell if click fraud is going on.
Last edited by cyrix; 10-07-2010 at 07:10 PM.
Looking for a programmer and designer? PHP\MySQL\C#\VB.Net\Javascript - Hit me up!
Somebody with three posts has a plan! Yes, there is gold in them dar clicks! Only problem it's guarded!

They have statistic about ip address n country, user agent, referrer, how long visitor stay etc. Don't forget, some proxy are detectable, and some ip are marked as proxy.
they read your posts and know all the methods
I don't want play fraud with uncle G
I want to know only what method Google use for click fraud![]()
click fraud? did you mean go to the local library and click your own ads?
Click fraud might be detected in many ways. As suggested already there are checks in place that include the overall CTR, your IP, referrer, time spent on a page, UserAgent by means of specific probes that do not rely on what the browser "claims" to be in the server env variable, cookies, flash LSO, recently silverlight and the likes.
To better the odds of identifying your "uniqueness" there might be checks on your monitor resolution, pixel depth, plugin capabilities and the ubiquitous click motion javascript bits to determine if the mouse movements are likely "human" or not. Moreover it's easy by means of a few lines of js to test which websites you visited already (css color trick) so that the likelihood of identifying two people with the SAME values becomes very thin...
These are all valid techniques to generate the so-called evercookies. It's also evident that a LSO cookie works seamlessly across different browsers on the same machine. With that being said (there would be many other things to say...) G is not god as many people here -that evidently don't know what the hell are talking about- claim. And -of course- can be tricked. Many of these statements about G remind me a lot when Oracle started advertising their DBs as "Unbreakable"... LOL. As always the great Sheep Effect ensues...
Of course G has world-class engineers and tricking them is not THAT simple that every noob fresh of understanding a few lines of js and what cookie and a session are can do it, right away, on command... With that being said, it's far from impossible.
Last edited by stradivari; 10-07-2010 at 11:27 PM.
promosirupiah (10-07-2010), Sprouts (10-09-2010)
Great use of analytics (google) from reffering sites, how genuine is your organic search: say you have a bot that writes "sex" and gets the 3rd result clicked" google sees that you just search "sex" and enter site. (just example),
Google bots are maybe the closest to artificial intelligence in the world.
I recommend not to do click fraud and everything would be fine! Common guys!
Based on referers, cookies and mostly "BEHAVIORS" (high bounce rates)
ฬ๏гк ฬเรє
basically what's been said here..
they look at all the available information channels
then they do various advanced statistical analysis to determine if anything looks suspicious
if any flags are raised, bring in an actual human to look at the data and your site
kinda similar to tax fraud detection in a way
one thing you can't really fake is that google can pretty much track the exact paths of everyone that has various google software installed, like toolbars and crap. they'll also make a note if the IP is currently logged into their gmail account. if none of the traffic you're sending has any of these, it'll get flagged for sure.
and if you were to go out of your way to randomize google accounts and google software for every single visitor you were sending, it would just be a ton of work and not worth it at all.
imo, the best way to get away with fake clicks is to mix them in with real clicks at a very high ratio of real clicks to fake clicks. gotta hide within the randomness that's already there. if you try to create your own randomness, you will almost certainly fail horribly as creating realistic randomness is an extremely difficult task.
We are what we repeatedly do. Excellence, then, is not an act, but a habit. ~Aristotle
They are detected in numerous of ways, most of which is baffling to an above average internet user. Click fraud can be done, if you know what you are doing. And of which 99% of everyone will not know and will be caught eventually.
Most fraud in any given level is detected primarily through IP Addresses and Cookies. As things get more advanced so does the tracking. Many of you guys probably don't know what LSOs are, but they are used intensively in all major sites.
There are a couple of ways advertisers and sites track you. Majority uses cookies and IP addresses, and some use LSO or flash cookies. Oops and lets not forget browser footprints. Browser footprints are going to be the new trackers as they are usually unique as well, about 1:1+ million for most people, unless youve just reinstalled your O/S.
Heres the breakdown of tracking internet users:
Majority use:
IP Address
Cookies
LSO / Flash Cookies
GEO IP
Then they can use more advanced tracking methods if it has to do with ecommerce or larger sites such as ebay/amazon/paypal/payment processors etc...
Email domain
Your Local Time
Browser Footprint/Fingerprint (Useragent/Browser type/version, Date/Time, MIME Type, HTTP Headers, .NET framework headers, lang, all your PLUGINS) This info can be very unique
O/S
Your System Fonts (Big footprint as well)
Helper components
Screen and resolution info
Your Bandwidth
CPU
CSS history exploit
Meta Data
and plenty more I can't seem to get off the top of my head. This is what you need to change.
As for click fraud, add in
Referrer
Length of Visit
CTR
Visits per page
and more analytic stuff
XRobot ? ....mmmm You like to cheating PPC Companies right ? be honest man cause ME TOO lol hahahahaha.
To The Point & it's Bonus For you. You can Cheating :
CPM ads, I recommend Cpxinteractive,Snoork, and Burstmedia
For CPC ads, you may try Adbrite,Chitika,Tribal Fusion, and AdsDaq
I can get easy money from them, that's mean you too.
Don't ask HOW, Just think. I am not Teaching Kevin Mitnick ways here.
Good Luck !![]()


IMO -
Traditionally:
Referrer
IP (that means more than it sounds like)
Cookies
Browser (and add-on/plug ins)
Mouse movement (for frames/laptop/desktop)
OS
Language
Your sex (as determined by 'the' variables)
Your age (as determined by 'the' variables)
Additional factors are those that can be sorted in AdWords campaigne.
There are the variables that I know of, there are approx. 30 variables that are transferred with each click and I do not know most of them. These variables create a "user ID" and when "user ID's" re-occur it provocatively raises a red flag.
Put a cookie (g analytics) on a forum (this one sometimes), see who is asking how click fraud is detected, keep his IP on file, compare that IP to the IPs of the latest suspected partners of adsense, ban the match.
Takes less than a second to run the procedure on a 200 MHz PC.
Bookmarks