Everybody knows that Google Inc.'s (GOOG ) innovations in search technology made it the No. 1 search engine. But Google didn't make money until it started auctioning ads that appear alongside the search results. Advertising today accounts for 99% of the revenue of a company whose market capitalization now tops $100 billion
Now, research is showing that Google's auction methodology, invented internally and so important for its success, is far more innovative than auction experts once believed. While superficially similar to earlier types of auctions, it is a "novel mechanism" that "emerged in the wild," write the authors of The High Price of Internet Keyword Auctions, a new study by Benjamin Edelman of Harvard University, Michael Ostrovsky of Stanford University, and Michael Schwarz of the University of California at Berkeley. Google's AdWords became so successful after its debut four years ago that some of its key features were quickly adopted by Yahoo! Inc. (YHOO ), then the search-ad leader.
MATHEMATICAL RIGOR
Close-mouthed Google has opened up about AdWords since the three economists cracked its code last November. It freed Hal R. Varian, a Berkeley economist who consults for Google, to publish some of his findings about the auction methodology. And on Feb. 22, Google gave an interview to BusinessWeek in which for the first time it named the technical leader of the project: Eric Veach, a veteran of Pixar Animation Studios (PIXR ) whose Stanford doctorate was in computer graphics, not economics. "Without his mathematical rigor we wouldn't have been able to do it," said vice-president of product management Salar Kamangar, himself a biology major, who was Employee No. 9 at Google and led the nontechnical side of the project.
Some of Google's innovations are only now being matched. For instance, Yahoo gives the top spot on its search results page to the advertiser who pays the most per click. But Google maximizes the revenue it gets from that precious real estate by giving its best position to the advertiser who is likely to pay Google the most in total, based on the price per click multiplied by Google's estimate of the likelihood that someone will actually click on the ad. Anil Kamath, chief technology officer of Efficient Frontier Inc., a search-engine marketing firm in Mountain View, Calif., estimates that Google earns about 30% more revenue per ad impression than Yahoo does. Kamath says Yahoo is likely to follow Google's lead soon. Asked about that, a Yahoo spokesperson says the company is "currently evaluating" making more use of the "click-through rate" in placing ads. Last fall, Microsoft Corp.'s (MSFT ) MSN embraced Google's approach, tweaking it to increase ads' relevance, when it began auctioning search ad space.
What makes Google's auction so different? Auctions come in two main flavors. In a typical first-price auction, participants put in sealed bids, then the winner pays his or her bid. But the danger is the high-bidder ends up regretting having won, an effect known as the winner's curse. A second-price auction lessens winner's curse because the highest bidder gets the prize but pays only the minimum necessary to win, namely the second-highest bid, plus perhaps a penny.
Kamangar, Veach, and colleagues chose a second-price auction. But not knowing theory, they designed one that differed in a key respect from the one economists had studied. In the economists' version, bidders always have the incentive to tell the truth. In Google's auction they don't, say Edelman, Ostrovsky, and Schwarz, since in some cases, by understating the top price they're willing to pay, advertisers could get a slightly lower position on the search page for a lot less money. They conclude that naive advertisers who told the truth could overbid. Google's system has pluses for advertisers, too, says Varian. It's easier to understand than the academic version. And it's proven to work on a large scale.
AdWords Select, as it was called at its February, 2002, debut, was actually Google's third crack at an ad auction. The first two were flawed, but Google founders Larry Page and Sergey Brin kept pushing. Even the current system isn't perfect. Advertisers complain that it's too much of a "black box." Still, if the best measure of innovation is commercial success, Google's AdWords was a grand slam. Says Kamangar: "Third time's a charm."
Now, research is showing that Google's auction methodology, invented internally and so important for its success, is far more innovative than auction experts once believed. While superficially similar to earlier types of auctions, it is a "novel mechanism" that "emerged in the wild," write the authors of The High Price of Internet Keyword Auctions, a new study by Benjamin Edelman of Harvard University, Michael Ostrovsky of Stanford University, and Michael Schwarz of the University of California at Berkeley. Google's AdWords became so successful after its debut four years ago that some of its key features were quickly adopted by Yahoo! Inc. (YHOO ), then the search-ad leader.
MATHEMATICAL RIGOR
Close-mouthed Google has opened up about AdWords since the three economists cracked its code last November. It freed Hal R. Varian, a Berkeley economist who consults for Google, to publish some of his findings about the auction methodology. And on Feb. 22, Google gave an interview to BusinessWeek in which for the first time it named the technical leader of the project: Eric Veach, a veteran of Pixar Animation Studios (PIXR ) whose Stanford doctorate was in computer graphics, not economics. "Without his mathematical rigor we wouldn't have been able to do it," said vice-president of product management Salar Kamangar, himself a biology major, who was Employee No. 9 at Google and led the nontechnical side of the project.
Some of Google's innovations are only now being matched. For instance, Yahoo gives the top spot on its search results page to the advertiser who pays the most per click. But Google maximizes the revenue it gets from that precious real estate by giving its best position to the advertiser who is likely to pay Google the most in total, based on the price per click multiplied by Google's estimate of the likelihood that someone will actually click on the ad. Anil Kamath, chief technology officer of Efficient Frontier Inc., a search-engine marketing firm in Mountain View, Calif., estimates that Google earns about 30% more revenue per ad impression than Yahoo does. Kamath says Yahoo is likely to follow Google's lead soon. Asked about that, a Yahoo spokesperson says the company is "currently evaluating" making more use of the "click-through rate" in placing ads. Last fall, Microsoft Corp.'s (MSFT ) MSN embraced Google's approach, tweaking it to increase ads' relevance, when it began auctioning search ad space.
What makes Google's auction so different? Auctions come in two main flavors. In a typical first-price auction, participants put in sealed bids, then the winner pays his or her bid. But the danger is the high-bidder ends up regretting having won, an effect known as the winner's curse. A second-price auction lessens winner's curse because the highest bidder gets the prize but pays only the minimum necessary to win, namely the second-highest bid, plus perhaps a penny.
Kamangar, Veach, and colleagues chose a second-price auction. But not knowing theory, they designed one that differed in a key respect from the one economists had studied. In the economists' version, bidders always have the incentive to tell the truth. In Google's auction they don't, say Edelman, Ostrovsky, and Schwarz, since in some cases, by understating the top price they're willing to pay, advertisers could get a slightly lower position on the search page for a lot less money. They conclude that naive advertisers who told the truth could overbid. Google's system has pluses for advertisers, too, says Varian. It's easier to understand than the academic version. And it's proven to work on a large scale.
AdWords Select, as it was called at its February, 2002, debut, was actually Google's third crack at an ad auction. The first two were flawed, but Google founders Larry Page and Sergey Brin kept pushing. Even the current system isn't perfect. Advertisers complain that it's too much of a "black box." Still, if the best measure of innovation is commercial success, Google's AdWords was a grand slam. Says Kamangar: "Third time's a charm."
0 comments:
Post a Comment