Google's recent announcement to offer CPA advertising has gotten a lot of attention. On my last entry on this topic. I got an interesting comment asking about accurate tracking and click fraud possibilities.

Tracking:
It will be interesting to see how they offer these types of CPA offers for sale to advertisers. My guess for tracking results is that they will use the conversion tracking technology that they provide as part of AdWords now, where the publisher will be required to place some sort of conversion code on their thank you for ordering page.

Assuming the initial deals are cost per lead, the advertiser would supposedly set an amount they are willing to pay for a lead, and when that confirmation page loads, Google will track this as a lead.

Here are some potential fraud scenarios:

Fake Leads / CPA Fraud:
Obviously there could be a scenario similar to click fraud, where instead of people around the world being paid to click on ads, they would be paid to generate fake leads, thereby generating revenues for Google and the sites hosting these CPA offer ads. I am sure this is a scenario that Google has thought of, and has looked at ways to track and prevent this, it will be much easier for the advertisers to identify fraudulent leads than fraudulent clicks. Google will need to come up with a way for advertisers to report these fraudulent leads. If they are simply using the conversion tracking codes, they will also need to deal with duplicate leads and page reloads the conversion pages.

Broken Site / Leakage Issues
The second type of "fraud" could come from the publisher's end where they don't always present the Google conversion code for all transactions, and thereby get clicks for free to their site without paying for "conversions".

Imagine this scenario. Someone creates a page to capture mortgage leads, and throws a lot of AdSense pay per click ads on their page, and due to some "technical glitch" the customer have difficulties completing the lead form, all they can do is click on the AdSense ads or leave the site. The site owner would then be able to get clicks for free because they would have little to no conversion rates. An advertiser might be willing to pay a lot per conversion if only 1-2% of the clicks actually convert on the page, but the non-converters drive revenues.

This might not be done intentional; there are a lot of sites out there with horrible shopping cards and conversion flows. Additionally many site have multiple revenue sources, and there will be a lot of scenarios that the CPA deal that is tracked is just one of the ways the site is making money, Google and their publishers might not share in a lot of the revenues, and thus allow advertisers to pay less per click for their traffic.

By moving into the CPA deals, Google will be taking on more risk in the transactions. There PPC model is a lot simpler to operate and administer from Google's end, because there is no transaction risk. In this type of free market, advertisers are willing to pay to the point of not being profitable, and in a perfect world, Google should be maximizing the return in the PPC model.

I guess my point for Google is, if it isn’t broken, don't fix it. This transaction risk is what the advertisers and affiliates have been arbitraging on Google for some time using their PPC advertising model. When ads don't work, we bid less or turn them off, as more and more bidders come into the market, Google's PPC model seems like the best and simplest way for them to make money.

My guess is that Google will find the CPA world offers more headaches than it's worth.

But you never know, if they ad a lot of intelligence behind the CPA model like they did with the PPC model, then only the best converting and most money generating ads (for Google) will display throughout the network, and the could open up a whole lot of new revenue sources and put some of these fly by night “CPA networks” out of business.