|Got my Client's Settlement|
They're not going to be happy
Got click settlement in client's account, on a fairly heavy advertiser - in a high fraud target zone.
About a half a day's clicks. They spend around $1 million a year. They are not going to be happy.
Another win for Google.
And the fraud will continue as a cost of doing business. The more competitive your word space, the more you will have to figure in for theft (shrinkage).
Under what circumstances would your client stop advertising with G, period? It's only when enough advertisers JUST SAY NO that anything will change (outside of a federal investigation) ...
Or maybe some geek will come along with a smart code to stop click fraud. Google says they only have, what, about 30 odd people looking for it or dealing with it? I seriously doubt the intellectual resources are being put into solving this problem, instead of just looking for fraud. When you consider how much money is being lost by advertisers, Google's paltry manpower allocation speaks volumes about its commitment (lack thereof). What is this? Thirty-six people out of how many thousands of Google employees?
|Google said it works with local law enforcement to shut down click-fraud rings. Shuman Ghosemajumder, Google's product manager for trust and safety, said the firm employs about three dozen people to monitor click fraud, 20 of whom respond to advertisers who report anomalies on their sites. Google keeps lists of hundreds of sites associated with pay-to-click networks, most of which he said target less-known search engines with weaker security controls. Google said it issues refund credits to advertisers when its technology detects irregular click activity, but it would not disclose how much is refunded. [Source: [msnbc.msn.com...] |
This is all reactive, not proactive. Necessity is the mother of invention, but so far no invention, probably because Google doesn't consider it a necessity. For all the brilliant minds which supposedly work for Google, they should be able to solve this problem.
IMHO, it's entirely *pro*active for G to *in advance* discover the IP addresses of click networks so that G can presumably ignore clicks from those networks if/when they happen without having to try and guess if they were "bad" or not.
Heck, I screen some compromised machines (eg in the SPAMHAUS xbl list) from certain activity on my site. I call that pre-screening *pro*active. Are you saying that I have to find the individuals concerned and lock them *and their children* up (in case they have bad genes) and throw away the key to be *pro*active rather than just boring old *re*active? B^>
Not that reactive is bad if it proves (as) effective.
Click fraud is to online advertising as shoplifting is to retail sales. It is a cost of doing business.
In both cases you try and stop the most egregious cases to keep it from becoming commonplace. But at a certain point the policing actions start costing more than the amount lost. In most cases this point is reached at a very small percentage of the losses.
If an advertiser's campaign is successful even with some fraud, they would be foolish to stop the campaign just to try and teach Google a lesson.
|This is all reactive, not proactive. Necessity is the mother of invention, but so far no invention, probably because Google doesn't consider it a necessity. For all the brilliant minds which supposedly work for Google, they should be able to solve this problem. |
Given the current Internet architecture, protocols, and service model, it is impossible to stop click fraud. To detect click fraud reliably, one must be able to determine the intent of any click. Intent is not communicated in a click.
It is possible to stop some kinds of click fraud, but what's really happening is some people have decided that certain usage patterns are "invalid" or "fraudulent" (such as some # of clicks from a (small) set of IP/useragent pairs over some (small) time period). A savvy fraudster can easily circumvent these defenses.
|IMHO, it's entirely *pro*active for G to *in advance* discover the IP addresses of click networks so that G can presumably ignore clicks from those networks if/when they happen without having to try and guess if they were "bad" or not. |
A problem with this approach is that IP networks are not fixed usage entities. If someone knows of a click network and screens it, then the fraudsters give up and/or move somewhere else, one could very well be blocking legitimate traffic. Another problem is that IP networks can change hands, so a compromised network may no longer be compromised once it's transferred to a new organization.
Google have a little-known method which totally stops click fraud affecting your Adwords account.
Disable Adwords click fraud [adwords.google.com]