Why CPA is Not a Cure for Click Fraud

Rarely does the first solution to a problem actually succeed. How often do we hear about a new breakthrough drug that's supposed to work wonders? Then comes a realization: oops, this doesn't work as well as it was expected to, and now we need a new approach. This seems equally true in the marketing biz: when a new technology or advertising program is rolled out, we can't be sure that it will appeal to its target audience or solve the problems at hand. We've really just got to watch and wait.

Just like everyone else, trial and error is sometimes our only choice.

This should be easy to see in the rapidly-changing debate over click fraud. Of course, click fraud has been around for a long time, but in the last year has a lot of serious attention been paid to overcoming the problem. The discussion has calmed down for the moment, but you never know when it might flare up again. Moreover, I think we might learn a lot about future initiatives by taking a closer look at what's been said in the past.

There have been three major suggestions for dealing with click-fraud:

  1. Look to the FTC: This hasn't been mentioned recently, but about a year ago -- when reports of rampant click fraud began to surface -- it wasn't clear at all what the response should be. Google, approaching its IPO, was talking up its "fraud squads" but many wondered if click fraud would prove to be a serious problem for search-based advertising.  So it was suggested by some that the FTC needed to step in and mandate changes.
  2. Take a case-by-case approach: Cooler heads -- primary amongst them, iMedia's own Kevin Ryan -- prevailed. Early and overblown reports of click fraud were scaled down, and instead of looking for an industry-wide answer, it was suggested that marketers should keep a closer eye on where clicks are coming from. Ryan outlined a number of methods for tracking ads that might be fraudulently clicked on, and new technological solutions began to appear from companies like ClickLab, and Zunch, which has a new Click Fraud Detective.
  3. Move away from CPC: More recently, Ron Belanger argued here in iMedia Connection that to deal effectively with click fraud many of the smaller search engines need to move away from the cost-per-click payment model, and instead embrace the cost-per-action (CPA) model used by many affiliate networks. The affiliate networks seem to have done well with it CPA, and many of them are moving away from CPC entirely. And they, as Belanger suggests, will tell you that CPA is something of a panacea for most issues of fraud.

This newest argument has much to recommend it. And I think my basic inclination is to agree with Belanger: CPA avoids click fraud. So, if the search engines were to make this move, there would be less click fraud. There's something to this. But it forgets that for this to occur, search engines are going to have to make the move. And this, I think, is less obvious.

Most of the affiliate networks that I've spoken to recently generally confirmed Mr. Belanger's belief that CPA is a remedy for Fraud. Elizabeth Cholawsky, Vice President for Marketing at ValueClick -- the parent company of affiliate leader Commission Junction -- said that ever since 2001 Commission Junction has only worked with a CPA model, in large part to avoid problems of click-fraud.

But Ms. Cholawsky also pointed out that CPA doesn't get rid of fraud entirely. "We have a dozen people whose job is entirely network quality," she said, and I got that sense that she felt it wouldn't be bad to have even more. There hasn't been a lot of chatter about CPA fraud (although there are a few reports), but this shouldn't necessary lead to treating CPA as a true panacea for fraud.

Others in the affiliate space agree. Choots Humphries, President of LinkConnector advised me that "there's still a lot of fraud in affiliate marketing." CPA pricing goes a long way, he argued, but one still has to be on guard. And just as important, Mr. Humphries said, was the fact that "a lot of merchants left the affiliate game because they couldn't participate in CPA" due to the increased technological sophistication necessary.

The real difficulty may actually have more to do with incorporating search engines and merchants into the CPA model than avoiding click fraud once they're there. When I asked Lance Podell, CEO of Kanoodle -- a search engine mentioned explicitly by Belanger -- if his company had any plans to move away from the CPC pricing model, he told me, "Kanoodle's current pricing standard is cost-per-click, and while we are always looking for improvements to our business, we do not have plans to change this model in the foreseeable future."

I think we can understand why. There's a certain niche to be filled here -- a lot of smaller merchants choose CPC because of its simplicity and perceived efficacy. And if the search engines were to shift towards the more complicated CPA model, they could easily alienate a large number of these merchants.

The pay-per-click model is probably here to stay -- and not just with the biggest players like Google and MSN. It might make sense for some players to move towards the CPA model eventually, and we should always applaud an idea that looks for a way that the industry can police itself. But for the moment, we're probably going to have to stick with the more incremental solutions, dealing with cases of click fraud individually (and there are a number of new technologies that are making this more effective). And even moving to CPA doesn't entirely alleviate the need for a bottom up approach -- pay-per-click or not, you've still got to watch those traffic numbers.

Isaac Scarborough is manager of market intelligence at Chapell & Associates. Read full bio.