|Clickbots and Adsense Fraudulent Clicks|
| 12:13 pm on Aug 6, 2008 (gmt 0)|
Ok, I would not do this but whats to stop me from writing a script to skip randomly across anon proxies queering google for certain search terms and then following the google result links to my own sites and clicking through adsense adds at a random clickthrough rate thats somewhere close to realistic and a second script to just create illusion of traffic to site in same manner to keep things in proportion?
Any stories of click through fraud here, either succesful or not?
I'm just interested thats all.
| 12:25 pm on Aug 6, 2008 (gmt 0)|
Google is pretty sophisticated these days.
I am sure some click fraud is successful, at least for a while, but unless you have some pretty neat ideas, chances are it's been tried before, and has already been rumbled.
The risk will also vary with which country you are in; where Google is well established, a prosecution is probably more likely. Where they cannot do that, I suspect they operate much closer to zero tolerance, and would be crazy to give you the benfit of the doubt.
But your guess is probably as good as mine :)
| 12:59 pm on Aug 6, 2008 (gmt 0)|
|writing a script to skip randomly across anon proxies |
I have no idea about any others but I do know that the AdSense Preview sites I use whilst showing ads state:
|The Adwords advertisers will be not be charged for the ad impressions or clicks generated through this tool. |
| 1:17 pm on Aug 6, 2008 (gmt 0)|
how many open proxies are there in the wild? How many of these does Google know about?
| 1:26 pm on Aug 6, 2008 (gmt 0)|
You have no idea how big Google adsense tracker matrix is... Almost, all Google properties includes tracking and everything goes to central system. They also purchase data from 3rd party such as ISP and all those so called 3rd party free statistics programs.
You don't search using Google, Google does search you ;)
| 2:44 pm on Aug 6, 2008 (gmt 0)|
You'd have to mix it in pretty well with actual valid queries in order to keep it under the radar - Google is all about finding patterns of misbehavior.
| 4:48 pm on Aug 6, 2008 (gmt 0)|
>>>at a random clickthrough rate thats somewhere close to realistic...
Isn't a realistic rate what it was receiving before dropping the clickbot into the site? To me that appears to be the flag waving at Google's algo.
- How would you match your conception of realistic to what Google knows is realistic from historical search trends and visitor activity to your site and sites like it? Dropping a clickbot to click on SERPs creates a change in those trends.
- How realistic is an extraordinary surge of queries?
- How realistic is an extraordinary surge of queries matched to visitor profiles that do not conform with the historic average of combinations of browsers and ip ranges?
- How would you accurately spoof statistical averages for traffic that is coming from Comcast, AOL, British Telecom, etc. combined with spoofing a realistic statistical average range of browsers and OS's etc. and match those to Google's data of what constitutes realistic?
| 5:00 am on Aug 7, 2008 (gmt 0)|
Ok, I get your point but these are still things that could be designed into a script i.e different se's, user agents, gradual increase of traffic over time etc.
| 6:57 am on Aug 7, 2008 (gmt 0)|
I'm not convinced you get the point. I'm not saying you are wrong. In fact, we're just throwing around opinions and I appreciate you starting this discussion. Just saying I'm not convinced the average spammer can build an effective clickbot that won't get caught. The theoretical clickbotter is up against a massive amount of Google data that can be used to spot and nuke them.
|...but these are still things that could be designed into a script... |
That's debatable, and here is why. You don't have the massive amount data to accurately model the statistical/historical occurences of specific browsers, IPs, referrer data, and more. Google is working with data direct from ISPs, their own search engine, the AdWords program, as well as data culled from GA and other sources. Massive amounts of data upon which they can create statistical models for what is normal and what is not.
That's not to say it can't be done. I imagine there are some who can probably get it to work, but I suspect the industrious clickbotter will have to obtain a lot of log files in related niches and calculate the averages. However, this is the easy part. It gets harder.
|gradual increase of traffic over time |
Across the entire Internet? This is, I suspect, the hard part.
- How many keyword queries should the clickbotter plan to inflate? Ten? Twenty? A hundred? A thousand? Do you realize how many longtail permutations there are within niche buckets?
- How will the clickbotter distribute these queries across the SERPs? What would be the precise historical ratio for distributing queries to the Ads, then to the Organic serps? A certain percentage of clicks go to the adwords ads in the Google SERPs. The rest is divvied up to the top three positions of the SERPs. Beyond position four what ratio will be used to distribute the rest? What position is the clickbotters site? On the second page? Do you realize how many artificial queries you're going to need to do inflate SERP clickthroughs so that you receive a "reasonable" amount of traffic, while distributing the rest of the clicks to everyone else (not to mention fake clicks on their sites)? Otherwise it's not going to look normal?
- Is your bot going to clickthrough adsense ads on other sites, too, to make it all look normal? That way it's not only the clickbotters site that's suddenly getting clicks from proxy IP addresses, to keep it normal and scaled?
- What about conversion statistics? Won't Google notice a dramatic increase in queries, followed by clicks, followed by a significant drop in conversions?
This is what I can think of that the clickbotter and all the other scammers are up against, inlcuding the ones riding their bicycle to thirty different internet cafes and clicking ads. This is what the click farms are up against when they're buddy clicking each other's ads. That's why people show up here lying about being innocent about getting banned and asking if they could reopen the account in a "brother's" name, etc.
[edited by: martinibuster at 9:12 am (utc) on Aug. 7, 2008]
| 9:06 am on Aug 7, 2008 (gmt 0)|
This whole business reminds me of a conversation from "the Dark Knight":
|So, let me get this straight, you think that Gotham City's richest, most powerful and ruthless businessman spends his spare time as the most effective crime fighter in the city, tracking down and beating up criminals. And you are intending to blackmail him?* |
|So, you recognise that Google is the most powerful force on the internet today, with an army of PHDs, and a record of suspending accounts and closing them, without compensation, notice or appeal. They've proved to the nth degree how sophisticated their systems are. And you think you and your pals can get away with a 2003-style click fraud? |
*apologies for errors in memory!