Google and CPA

News surfaced a little more than a year ago about the beginnings of a Google CPA network. In June of 2006, word started to break about a product called Content Referral where, for the first time, advertisers could pay on a per action basis rather than on a CPC basis. The announcement came at a good time for the company as it provided a respite from the ever increasing click fraud discussions dominating the landscape. In typical Google fashion, Content Referral started out invite only for both advertisers and publishers. Despite the incredible buzz and speculation it started, content referral and Google CPA quickly faded from the forefront of Google watchers. Nine months later, in March of this year, the topic resurfaced when content referral gave birth to the latest iteration, Google’s Pay-per Action program. Similar to the presumably defunct Content Referral, PPA also started out invite only, and instantly gained the nickname the Commission Junction killer.

This is not the first time outsiders have dubbed a Google product, the "xyz" killer. When Google unveiled their display ads product, with their now ubiquitous text ads in a banner format, it seemed that it would not only compete head to head with the traditional graphical display ad networks such as Advertising.com and Valueclick, but significantly diminish their business. Interestingly, Google’s entrance ended up acting like the tide that lifted all boats. Instead of taking away inventory from Advertising.com and Valueclick, Google ended up helping a super long tail of content monetize better, inventory that didn’t work with the display ad networks. For those running ads from the display networks, many simply added Google into the rotation because they had excess impressions or created entirely new ad space on their pages because Google ads blended into content well. In the end, more overall impressions became available for everyone, with Google gaining the lion’s share of this new revenue.

Usually, when Google launches a new program we gain insight by looking at the advertiser perspective. The publisher side though, often provides equal if not more perspective, and when available, it certainly helps complete the picture. When Google launched Content Referral, they didn’t make a mention on their advertiser focused blog, as was the case with the follow-up Pay-per Action program. Nor did Google discuss the program on their publisher focused blog. Thanks to Jennifer Slegg of the well regarded JenSense.com, we can get a peak into how Google pitched Content Referral as she obtained the email sent to qualifying publishers. In that note Google addresses, among other things, the hypothetical question, "Do these compete with regular content ads?" They answer saying, "These ads will not compete with contextually targeted ads. Instead, they will show across a separate network, the Content Referral network. To place one of these ads on your site, you can set up a new ad unit that supports any of our current ad unit sizes." These ad units are "separate and appeal to different types of users. These CPA ads are also additional inventory to your existing AFC (Ad Sense for Content) ad units."

When Pay-per Action first launched in March, many immediately assumed that, unlike Content Referral, the new round of CPA ads would appear on their search engine. Mirroring their Content Referral roll-out, that was not the case. As the official Adwords blog pointed out, Pay-per-action ads are only shown on sites in the Google content network and that "Publishers choose specific pay-per-action ads that are relevant to their site and can place them in a new ad unit on their page." This is an important point, and it’s the reason that people thought PPA and its predecessor might compete with true affiliate networks such as CJ. For more read Scott Jangro who explains, best of all, why Google CPA will not hurt CJ. Unlike Adsense for content, where a publisher simply pastes code and Google decides the ads, the CPA programs would have the publisher choosing the ad. More important for our analysis comes not from the advertiser perspective but again from the publisher perspective. Just this week, Google did what it should, but it goes counter to what they said. Instead of publishers choosing the ads (Content Referral and Pay per Action) and not competing with "existing AFC" (Content Referral), starting this month, whether qualifying publishers like it or not, Google "will enable cost-per-action (CPA) ads to compete in your AdSense for content ad units on a limited portion of your traffic, 5% or less of all ad impressions."

No other network, affiliate, display, or other could pull this off. Google not only has an enormous publisher network, but they have leverage. Other companies would have to beg and overpay to test their inventory. Not Google. They simply say "we will." And their rationale, "Your account has been selected because we believe that you’ll earn more by having these additional targeted, high-quality ads competing in your ad units." And, "Rather than generating earnings for a click or impression, you’ll be paid a larger amount for each conversion with these new ads." We don’t know about you, but something doesn’t smell quite right. Generally, when a company wants to test an account, they don’t do it because they think you will earn more. They do it because they think they will earn more. They selected you because your account didn’t earn them enough and they question your quality. In other words, it starts with 5%, but soon, they will run 95%. More importantly, you do not decide which ads; they do. And, while it might sound minor, make no mistake; Google starting to test CPA in content moves them one giant step towards not just the ad network but the arbitrage business. As we’ve mentioned before, but it bares repeating, learn from them. Instead of doing the obvious, such as inserting CPA ads on properties that would seem to cost them nothing, like Google.com, they use their publishers. Using the publishers, they gather data across disparate data points, with an indeterminate opportunity cost, as opposed to their main page that provides only one data point and has a much larger cost. Once they figure out the metrics then, if they want to, they can move up the chain. This small test might seem like just another publisher release, but this could truly end up being the smallest big step towards the future of CPA and Google.