Google CPA Ads Will Worsen Click Fraud
There is a lot of buzz around Google's recent beta testing of a cost-per-action (CPA) advertising program. The excitement is justified of course since Google has a strangle-hold on the cost-per-click advertising market making it a power to be reckoned with and one worth watching.
The program's beta status means that us little folks can't run any tests of our own, so I'm forced to speculate instead. In particular, I'm thinking that click fraud will get worse with and not better as some people suggest.
The CPA model is supposed to eliminate click fraud because advertisers don't pay until a particular action has taken place on a specific website. This means that automated and manual systems that are used to click on ads will no longer result in any cost to the advertiser. The problem is that there's nothing stopping fraudsters from building systems to automatically perform the necessary action especially when the action is just filling out a contact form.
In fact, I'd argue that since CPA amounts will be significantly higher than the typical CPC amount, it now becomes economical to use manual systems i.e. inexpensive labor from low-cost of living countries such as India, China, and the Philippines. That is, where it took 5 clicks under the CPC model to generate $1 it will now take just 1 click to generate $25. In addition, detecting the fraud will become harder because CPC fraud required multiple clicks to be lucrative making it somewhat easy to detect patterns. With CPA fraud, far fewer clicks are needed to generate the same income making patterns harder to detect.
The one trick I think that advertisers can use to their advantage is to select an action that requires information that is hard to fake such as matching credit card and addresses. This will, of course, work for eCommerce sites, but what can the rest of the CPA advertisers do?