Some Consequences of Pay-Per-Click’s Growing Pains

Pay-per-click advertising has grown up, but what sort of creature is it now? Some savvy people have come up with ways to game the system in order to line their own pockets; at least one of these ways (click fraud) hurts those who use the system correctly. Is it time for a new online advertising model? What does the future hold?

When pay-per-click advertising first got its start nearly a decade ago, some Web surfers reacted with shocked outrage and betrayal. How dare search engines accept money from businesses in exchange for placing their websites in prominent positions when searchers used particular keywords?! Others simply figured the scheme wouldn’t work. What business would pay that kind of money for such a small advertising blurb that didn’t even sit there all the time, in contrast to a banner ad? But GoTo.com, later renamed Overture and now owned by Yahoo!, proved the scoffers wrong. Pay-per-click advertising has grown big, and we’re all suffering from its growing pains.

By now, everyone knows how pay-per-click advertising works. Advertisers bid on keywords, stating the amount of money they’re willing to pay every time someone clicks on their ad, which pops up when a search engine user performs a search on a particular keyword. How much they bid, in comparison to the other bidders for that keyword, helps to determine what position their ad receives (and presumably how likely it is that a user will see their ad and click on it). That in itself would be fine, if there weren’t ways to game the system.

Click fraud happens whenever users click on a business’ ad with no intention of doing research on or purchasing anything from the company. Usually click fraud is committed either to harm the advertiser directly by causing him to go through his advertising budget too quickly, or to increase the amount of money an affiliate website gains from hosting ads that are part of a pay-per-click advertising campaign. Reputable industry analysts estimate that click fraud runs as high as 20 percent for certain keywords.

That’s not a typo. You do the math. By this estimate, for every five dollars you spend on pay-per-click advertising, you effectively throw one away. Is there a solution, short of taking it to the courts, as Click Defense chose to do with Google? Snap.com seems to think so – which is interesting, since the force behind them is the same man who started pay-per-click advertising in the first place.

Bill Gross was the pioneer behind GoTo.com’s “pay-per-click” advertising model. Yahoo! purchased the company in 2003, one year after search engine giant Google introduced AdWords, its own pay-per-click advertising service. One could say that this makes Bill Gross indirectly the cause of click fraud – in the same way that Thomas Edison is indirectly the cause of all the neon-lit advertising in Times Square. Gross, described by the Associated Press as a “serial entrepreneur,” believes he has an answer to click fraud. It’s so simple that a lot of people must be slapping their heads because they didn’t think of it first.

With Bill Gross’ new advertising model, instead of paying a cost-per-click, advertisers shell out a cost-per-action (CPA). Business owners do not pay every time a user clicks through from an ad to their website. Rather, they pay when the searcher completes a desired action, such as makes a purchase, fills out a form, performs a download, and so on. In this way, advertisers only pay for concrete results.

Best of all, the model eliminates click fraud. Remember, click fraud occurs when someone visits an advertiser’s website with no intention of doing anything, for underhanded financial purposes. Since the advertiser is no longer paying for those clicks, click fraud (at least as it is currently defined) can’t happen.

Gross is launching this new advertising model from Snap.com, which describes itself as “a next-generation search engine for broadband users.” Launched in October 2004, it has received some good press. It is not likely to topple Google any time soon, however. In the nine months or so since its launch, Snap processed about 16.4 million search requests. Compare that with Google, which processed 1.8 billion searches in a single month – and that only counts the searches originating in the U.S.

Snap.com has some other interesting technology going for it. It combines algorithmic search results with human click stream data. Snap calls this “behavioral ranking.” The idea is that sites at which people have received more value will rank higher in the results than other sites. Value can be measured in the form of good content (presumably gauged by users spending more time at the site) or completed transactions. In this way, it bears some similarity to Yahoo’s recent My Web 2.0 initiative, the difference being that it directly measures user behavior rather than simply letting users tag and share sites they found useful.

If anyone foresaw the possibility of click fraud when pay-per-click advertising first came out, he or she is no doubt enjoying a well-justified “I told you so” moment now. After all, it’s not often that a major company’s chief financial officer stands up and admits that a problem is so serious it could potentially threaten the company’s business model, as Google CFO George Reyes did in December 2004. But there are other consequences of the pay-per-click advertising model that likely few people foresaw. One of these is the domain name gold rush.

To be more accurate, the “gold rush” can be blamed on a concatenation of circumstances. The main factor, of course, is how easy it is to publish pay-per-click advertising links from Google and Yahoo!. All you need to do is put up a site, join the Google AdSense program (which only costs time, not money) and/or the corresponding Yahoo! program, and you’re literally in business.

The second factor is how inexpensive it is to purchase a domain name. VeriSign, which runs the .com and .net domain names, reported that one could be had for as little as $6 to $7 per year. At that price, a website hosting pay-per-click ads doesn’t need to get many people clicking through its ads in order to turn a profit. The third factor is the five-day “Add Grace Period,” during which new registrations can be deleted for a full refund.

This leads to a situation in which close to a quarter of a million domain names are registered every week for only a few days, in order to test the traffic potential of the names before they are discarded. Since the domain names only need to deliver a small amount of money per year to return their owner’s investment, a large number of names that do even slightly better than that can return a respectable income.

The domain name gold rush caused VeriSign to change the way it reports the number of active registrations it receives for its domain name business. VeriSign CEO Stratton Sclavos said it would reduce the size of reported registrations by about two percent. “At the end of Q4 we probably saw about a 40,000 to 50,000 name difference between the end of quarter number and what happened five days later, at the end of Q1 there was about a 150,000 name difference, and at the end of Q2 it was about a 700,000 name difference,” he said by way of explanation. “We know that at the end of any given week, five days later a substantial number of names that just got registered will get deleted out.”

Any internet company that is still around at the age of ten is considered “mature.” But what about business models? Is pay-per-click advertising all grown up? Is click fraud merely a glitch in the system that the search engines can fix with greater vigilance and filtering? Or are attempts to game the system, both with click fraud and the domain name speculation mentioned in the previous section, symptomatic of deeper problems? 

Bill Gross has a proven track record of successful ventures behind him, with GoTo.com and its sponsored link advertising model being just the best known. Advertisers and business owners want measurable results. With his new cost-per-action advertising model, that is exactly what they are paying for. If Snap.com follows the same path as his earlier successful internet company, it won’t beat Google. Rather, it will achieve its success by seeing the other major search engines adopt their own versions of its advertising model, and then being purchased by one of them – perhaps Google this time, instead of Yahoo!.

The domain name speculation centering on pay-per-click advertising may itself mature into something else. To anyone who wants to set up a website for the sake of promoting a business or a beloved hobby, the approach seems to put the cart before the horse. Indeed, any SEO will insist that you must build content first – and most businesses would not care to advertise other businesses on their own websites, particularly rivals. But someone setting up a hobbyist site might gain a great deal from posting the appropriate pay-per-click ads from Google or Yahoo!. And somebody trying to gain business by promoting him or herself as an expert in a particular field might give the idea a second glance. At the very least, it is something else to consider and evaluate as to the return on investment.

This of course brings up one of the joys and frustrations of the internet in general: it is ever changing. Just when you think you’ve grasped it, it acts like one of those water-filled toys that wiggles out of your grip the tighter you try to hold it. With pay-per-click advertising, just when we have grasped what it can do for businesses and those who offer it, it appears that it will morph into something else. 

Google+ Comments

Google+ Comments