Why Dr. Eric Schmidt (Google CEO) may be wrong and right about click fraud

[Bob Russell]

The value of Google advertising currently depends pretty heavily on the ratio of "real clicks" to false clicks from click fraud. If advertisers are paying money for falsely generated clicks, then they are likely to balk at continuing to pay advertising fees. This is a threat to Google's revenue generation as awareness of click fraud rises.

Dr. Eric Schmidt, Google CEO, and his team have come up with a couple of ways to propel advertising. The first solution is to move from a pay per click model to a cost per action model. The idea is that advertisers don't pay for every click. They pay for the bottom line result, such as "whenever a site visitor clicks on the ad on your site AND performs a specified action, such as generating a lead or purchasing a product." (This text is taken from a Google AdSense e-mail via SearchEngineJournal.)

This would seem to be an ideal alternative for advertisers. You have very little hope of such a system in magazine ads, for example. There was a failed attempt to barcode such print ads, and tie results to ad responses, but it required the reader to have a barcode reader, and swipe the barcode as part of their response mechanism. I guess at the time, that sounded reasonable!

The second reason that Google says advertisers should believe in their online advertising was discussed recently at an event held at Stanford University. Eric Schmidt argued that the fraudulent clicks will be integrated into the pricing anyway, so the market will solve the issue, and there is no need to worry about the future. Specifically, he said, "Eventually, the price that the advertiser is willing to pay for the conversion will decline, because the advertiser will realize that these are bad clicks, in other words, the value of the ad declines, so over some amount of time, the system is in-fact, self-correcting. In fact, there is a perfect economic solution which is to let it happen."

Makes sense, right? And if advertisement hosts (and therefore Google in partnership with them) takes the hit, there really is no reason for advertisers to worry. Surely all is well, and Google only has its competitors to worry about, not click fraud, right?

Not so fast. Don't believe for a moment that this will make all advertisers agree with him that click fraud is no longer an issue once the market forces adjust pricing. And don't believe, either, that the only reasons Google is working to eliminate click fraud is "...because it is a bad thing, because we don’t like it, because it does, at least for the short-term, create some problems before the advertiser sees it, we go ahead and try to detect it and eliminate it."

But despite this slight exaggeration of the market's power to solve click fraud, we'll see below that it might actually reveal a bit more of Eric Schmidt's genius after all. First, the problem with his argument.

Just because click fraud is accounted for at a global level, doesn't mean it solves the problem. For example, here are two issues that are not solved:

1) The estimated amount of click fraud may not be the actual level of click fraud.

If this is the case, there is uncertainty about what is actually being bought by advertisers. They may still be overpaying if their estimates are wrong, or even if they are accurate but not timely and the click fraud rates show a structural change. Such an event could, for example, be due to a new click fraud technology that allows mass undetected fraud.

2) Everyone adjusts their pricing across the market for overall levels of click fraud, but click fraud is highly unlikely to affect all advertisers equally.

Hot topics may be more likely to attract click fraud because of efficiency. Or, conversely, maybe hot topics and major advertisers may be avoided by click fraud schemes in order to minimize their risk of being caught and prosecuted. Whatever the strategy of the click fraud schemes, the very fact that they will probably target certain ads means that some advertisers will be hit harder. That means that any particular advertiser may not be getting very much for their money. (Or maybe they are lucky enough to be getting a great deal if they face few false clicks.)

So, if click fraud is not solved by the market, what is important about the click fraud arguments coming from Google? Why do I say Mr. Schmidt might well be expressing genius in his statements? Again, two points:

1) The obvious advantage is that his statement about market corrections may help calm people's fears about paying too much for advertising. It turns the whole issue of whether click fraud can kill online ads paid for by clicks on its head, and replaces it with an argument about what the correct price shoul be.

2) While it may be true that the market can't fully correct for click ad fraud, there is another subtle point that may slip by the casual observer. Following Schmidt's argument, you still find that while the market adjustment only solve s the global price levels to a certain degree, it still changes the nature of the problem in a whole different way.

Specifically, it's no longer necessary to identify each false click. Now it's only necessary to identify the amount and locations of click fraud, not to stop it entirely. This is still a hard problem, but should be one less degree of difficulty. And it doesn't have to even be completely precise to be effective.

Aha, now it's starting to make sense isn't it? Google still wants to find and stop click fraud. Yes. But not because the Google engineers are just so very curious and find the whole thing interesting as a way to stop evil. In fact, it's at the core of the problem they still need to solve... how to identify patterns and levels of click fraud. To be able to estimate it for given advertisers so they know they aren't paying for worthless false clicks.

And also, of course, to minimize the click fraud in the first place, because it helps elevate revenues and advertiser confidence. Not as much as you might think, though, because all it really does is stop bringing in click events that advertisers shouldn't be paying for in the first place. Remarkably, click fraud is only bad for Google when advertisers avoid buying advertisements because of it. It's more of an advertiser confidence issue than a Google revenue-stealing issue.

So what's the bottom line? Google pretty much has things under control:
* Google is trying to limit click fraud, but as long as advertisers believe that have a handle on the percentage of false clicks that apply specifically to them, ad sales should stay strong.
* Google only has to convince advertisers that there is limited uncertainty about the levels of click fraud, not that click fraud doesn't happen
* If all else fails, there is a potential cost per action model waiting in the wings, and already being tested.

The tech and investment community may be increasingly suspicious of Google because they are so influential and successful, and may wonder how Microsoft will try to "head them off at the pass", and may think that there are still a lot of upcoming obstacles for the company to conquer. That may be right. But one seems to constantly be impressed at the level of talent they are building in-house, and the continual successes and well-executed business plans so far... even if they are winging it as they go!

From http://blogs.zdnet.com/micro-markets...back.php?p=219.