I give Peter Osnos credit for not being as nutty as Richard Posner or as self-pitying as Dana Milbank; his piece from CJR on "WhatÃƒÆ’Â¢ÃƒÂ¢”Å¡Â¬ÃƒÂ¢”Å¾Â¢s a Fair Share In the Age of Google?" (7-8/09) is the most reasonable version I've seen of the news industry's case against the search engine company. Still, I can't help but think that he's missing the point in a fundamental way.
One of Osnos' key examples of the unfairness of Google involves Sports Illustrated's website, SI.com, and a story it ran (2/7/09) on pitcher Alex Rodriguez testing positive for steroids. Osnos relates SI.com's grievance: Though it broke the story, other websites got as much or more traffic from it:
Most galling was that the Huffington Post's use of an Associated Press version of SI's report was initially tops on Google, which meant that it, and not SI.com, tended to be the place readers clicking through to get the gist of the breaking scandal would land.
From a journalist's perspective, this is patently unfair: SI.com got the scoop, and ought to get the reward. But is that the right perspective to look at what Google does? Journalists are not, after all, in the business of creating information; they're in the business of conveying information. Sports Illustrated's reporters did not create Rodriguez's failing steroid test results; Major League Baseball did that. People with access to the test information passed it on to SI, and SI put it up on the Web.
But that's not where the process of information transmission stops. People can't check every website that might break a news story of interest to them every day, so they rely on news gathering institutions to bring information together for them–that's what newspapers do, that's what AP does, and, yes, that's what Huffington Post does too.
Osnos attributes the Google results to the fact that "Huffington is effective at implementing search optimization techniques, which means that its manipulation of keywords, search terms, and the dynamics of Web protocol give it an advantage over others scrambling to be the place readers are sent by search engines." And it may well be that the folks at HuffPo are better at that stuff than SI is–though you'd think with the $84 billion entity of Time Warner behind them, the sports mag could afford to figure it out.
More important for HuffPo's search results, however, is the fact that people who use the Web have gotten used to looking for breaking news there, and so when they link to a story they find interesting they link to it there. Google's methodology, looking for links as a surrogate for how people use the Web, finds more of them going to Huffington Post than to SI.com–and that's why HuffPo came out on top.
Osnos says that "human help" needs to be incorporated into Google's algorithm–given that the search engine last year announced that it had indexed more than 1 trillion urls, this suggestion would seem to be rather impractical. But it's not clear that the human-free algorithm is making the wrong choice by directing Web surfers to the sites people most often go to when looking for information.