[ Pobierz całość w formacie PDF ]
.Brin and Page believed that people would trust machines, perhapsbecause they were familiar with the work of Joseph Weizenbaum.Weizenbaum, a founder of artificial intelligence (who later becameone its harshest critics, largely because so much research was financedby the Pentagon), was one of the first to highlight the singular andstrange relationship between man and machine or rather, computerprogram.In the early 1950s, he designed a computer program thatallowed a human being to converse with a machine.Much to hissurprise, he discovered that ordinary people  become emotionallyinvolved with the computer and.anthropomorphize it. 5 But evenif Page and Brin hadn t read Weizenbaum, they certainly worked withresearchers who had looked further into the relationship between34 Chapter 2human and machine and similarly discovered that  an individual sinteractions with computers and television sets are fundamentallysocial and natural, just like interactions in real life. 6 Many studieshave shown that the rules of reciprocity and courtesies that govern ourcontacts with friends are also used in our interactions with machines.Page and Brin set out to design a computer-based system thatwould create a comfortable and familiar environment and, in thecase of financial transactions, mimic the mechanisms people relyon to build confidence in a relationship: Specifically, a person fol-lows a learning curve, gradually engaging as he or she gains moreexperience.People are prudent, starting with small risks, and theyare most comfortable when they have a way to back out quickly incase of disappointment.The next section takes a meticulous look atGoogle s payment methods for advertising, which clearly demonstratethat programmers took human transactions as the starting point.Competitive BiddingWhen setting the payment structure for Google ads, following amodel like Yahoo! s and setting a standard cost-per-click wouldhave been simple.Instead, using Overture s example, Google chosea bidding system.Advertisers compete for keywords, and the morethey do, the higher the price of the keyword.Instead of using traditional ascending bids, as practiced in auc-tion houses, Google s leaders chose a system in which the bidderstates the maximum price he or she is ready to pay for a keyword.This price remains confidential, known only to Google.The sale ismade to the person making the highest bid, but at the next high-est price.The system encourages bidders to indicate the price theyare actually ready to pay because keeping it secret from the selleroffers no benefit; this system also prevents collusion because bidsare confidential.The AdWords bidding system resembles the method used forthe Google IPO, but the system differs in two important ways.First,bidding is continuous; the goal is to buy not a product, stock, orcontract, but a position on a screen that may change constantly.Thisencourages advertisers to experiment, vary their list of keywords andsettings, and correct their initial decision.They can, in other words,The Google Economic Model 35train themselves by tinkering.The gains from improvements can besignificant.As early as 2004, two Australian researchers (BrendanKitts and Benjamin Leblanc) demonstrated that modifications couldmultiply an ad s effectiveness by four times.7The bidding mechanism may appear complex at first glance, butmost of the complexity happens on the Google side.Advertisers areled through the process of setting a budget and their maximum bidprice for keywords; Google performs the calculations.By assigning advertisers higher positions based on both what theypay and the effectiveness of their ad (multiplying the payment priceby the click-through rate and ranking ads according to the result),Google encourages advertisers to invest time in learning to composeeffective ads, with both parties directly benefiting.The bidding process also offers many advantages to both thecustomer and Google.For one, bidding eliminates rate negotia-tions between advertisers and salespeople as well as concerns aboutprice increases, and it makes billing simple and more transparent.The customer decides how much to pay, and the market determinesthe price.Google has seen a direct benefit from the bidding process, too.In fact, the average price of Google s ads has probably risen higherthan it would have if Google had fixed its ad prices.According toadvertisers interviewed by MarketWatch in 2007, keyword searchprices on many terms had risen between 40 and 60 percent in 2006.8Of course, bidding success requires that advertisers spend timelearning the process and the rules of the game.But this time is notlost: Well-chosen keywords result in more clicks, more visitors toadvertiser websites, and more revenue for Google.No Content, No PortalHad they taken the advice of experts, Page and Brin would havebuilt a portal site with multiple services, just like their principalcompetitors.They chose to do just the opposite instead to produceno content on their own, focusing instead on offering tools to findor produce content.36 Chapter 2No Shortage of ContentAlthough Google doesn t generate its own content, the company veryactively develops and purchases tools to offer individuals ways to createweb-based content, whether that content is contained in a Blogger blog;an article in Knol, Google s  Wikipedia killer ; a website created withGoogle Sites; a video posted to YouTube; pictures displayed on Picasa;or code hosted by Google Code.And the list goes on.Thanks to thesefree tools, anyone can produce and publish content that attracts visitorsand offers more opportunities for Google to serve up ads [ Pobierz całość w formacie PDF ]