Click Fraud, Google, and the Federal Reserve

I've always been a little skeptical of all the uproar about internet click fraud.  Assuming an advertiser responsibly watches their ROI on internet ad spending, the price payed per click (which is set at auction) should adjust.  I'm actually not completely comfortable with the above statement (which I'll get to below), but apparently Google's CEO agrees:

Google CEO Eric Schmidt believes there is a “perfect economic solution” to click fraud: “let it happen.”

Schmidt discussed how the pay-per-click advertising model is inherently “self-correcting” in regards to click fraud during a Stanford University event last March. Schmidt extolled the enhanced trackability of the online pay per click advertising model versus pay per impression models, while acknowledging “smart but evil” people try to “go around system.”

According to Schmidt, Google’s auction-based pay-per-click advertising model is inherently self-correcting.

So in an efficient market, Schmidt and I see eye to eye.  But there's some obvious counter-evidence to this: people are making money for doing nothing.  So, say I publish a blog, run Adsense, and ring up a bunch of phony clicks (somehow evading Google's algorithm-based detectors).  Whether it's Google or the advertisers paying, someone's paying me for cheating.  That's a flaw.  It may be a tolerable flaw, but a system that rewards people for gaming it is not a perfect system.  If the advertisers don't mind (because they've adjusted their expenditure), then it's Google shareholders who are paying.

So let's say it is Google shareholders paying; maybe the company thinks the cost is worth it.  Perhaps the costs (direct or indirect) of aggressively combating would be higher than just paying the cheats.  That may be, and if so, then perhaps click fraud just isn't a fight worth fighting.

But even if this is Google's calculus (and this is 100% speculation, though I think the logic is sound), there's still a problem.  Click fraud, from month to month, isn't likely to be distributed evenly.  Attackers, in their quest to evade Google's watchful eyes, are likely to constantly change their tactics, attack different companies, try different kinds of content to run against ads (and thus get a different mix of ads).  In other words, I doubt that click fraud can really predicted well enough, so that an advertiser can feel comfortable with its ROI projections.  And I think that even if the numbers look steady in the big picture, it matters if inconsistency reigns on the micro level. 

Ok, so I mentioned the Federal Reserve in the title, so that needs to be tied in at some point.  There's a school of thought (which I would very much like to be sympathetic towards, cause it's so delightfully contrarian) that says the Fed activity is entirely irrelevant, that no matter what the Fed does, prices will just correct.  So, for example, if the Federal Reserve slashes rates hard, then the prices of goods like oil and housing will just rise in correction.  But while that looks like inflation, it's actually not a big deal, because cash will be more plentiful, to cancel the higher prices out.  If Eric Schmidt were the Fed Chief, he might say it's self-correcting.  But there's something I'm uncomfortable about in this scenario.  Yes, there may be a lot of extra cash sloshing about the system, but it's distribution isn't even, or even --gasp -- fair.  For example, not everyone has home equity to tap when interest rates are at 0%.  And price swings won't necessarily be across the board.  Ask the airlines if it makes a difference whether oil is at $35 or 75$.  In other words, while cash levels on the whole may rise to meet prices on the whole, there's going to inconsistent pockets of pain and pleasure.  And, taking this one step further, price inconsistency makes business planning hard, which seems bad for the economy.  And, bringing it home here, ROI inconsistency, when it comes to advertising online, can't be good for Google or its clients.

This is an interesting solution, and though I don't complete share Schmidt's sanguine view (though I'll concede he has a much better vantage point), I don't share the gloom and doom perspective of many, that click fraud is Google's (and the whole internet's) Achilles tendon.

Comments

[Trent]

Actually, it is neither Google nor the advertisers who suffer. Advertisers adjust down the price they are willing to pay per click, so that their total expenditure is comparable for what they would like to pay for only legitimate clicks. Google gets a discounted rate on each click, but is also paid for the fraudulent clicks so it comes out even as well.

So who pays? The honest content providers who get a lower cost per click but no advantage from the fraudulent ones.

Perhaps if Yahoo or MSN had a more efficient way of eluding fraud the honest content providers would shift to their ad systems to earn the higher payment per click that should result, which in turn would act as the enforcement arm to encourage Google to be more vigilant.

[toby]

I'll have to agree with Trent here, the problem is the same as the "lemons" problem. Unless you can correct information asymmetry, the market can collapse.

I wrote a piece about Adsense and lemons here:
http://blog.kiwitobes.com/77

[Reinhard]

Very intelligent comments by all of you, and I especially like the lemon story. I first totally agreed with Joe in the Google case, that the market adjusts and Google shareholders would be ultimately hurt. After reading your other comments, I agree that others could be hurt, too. All that Joe says is that someone must get hurt, since some people are "useless" sitting around and either clicking or programming robots, and that's not good for anyone! Somewhat agreeing with the previous poster, content providers and advertisers could both get hurt, simply because the spread goes up. (Same as when a penny stock has a bid/ask of $1.00/$1.30, and a market maker profits). And the liquidity drops (=Google's revenue). I totally agree with the Federal Reserve statement too, since I thought of that many times. Yes, the market can adjust, but it hasn't yet at the time the government makes the decision, and even after that it can't quickly. Say, they announce to double the money supply tomorrow. Savers will get hurt, borrowers rewarded. It is unlikely that we all had the same debt to income ratio, or even impossible since retirees and working people are just in an inherently inexchangeable situations.