It seems we spend a lot of time talking about lawsuits filed against Google these days, what with Sumner Redstone’s $1 billion swing at the Web giant and all.
But it’s a smaller suit just dismissed by Judge Jeremy Fogel of the U.S. District Court for the Northern District of California that may have larger implications and is infinitely more interesting.
KinderStart.com, a Norwalk, Connecticut-based company, claiming to be “the largest (and most popular) indexed directory and search engine focused on children zero to seven on the ‘net” sued Google in March 2006 alleging it had defamed the site by cutting it from its Web search ranking system. It argued that Google was behaving anticompetitively by skewing its search results to maintain its search industry dominance, reserving its top search placements for those paying premium fees.
But last Friday the judge quietly threw the case out, ruling:
“KinderStart had failed to explain how Google caused injury to it by a provably false statement … as distinguished from an unfavorable opinion about KinderStart.com’s importance.”
The ruling seems to be on solid legal ground, but here’s the thing: even if Google purposefully bumped a would-be competitor down in its rankings to maintain its global dominance, nobody would be able to prove it.
PageRank, Google’s trademarked process where a numeric value represents how important a page is on the Web, in only part of the secret sauce that makes the search engine so tasty.
Much like many recipes in the food world, a Google search is a matter of how much of each ingredient is being used — the “weighting” of each piece, reports Macworld. While there’s a bevy of information on the Web on the primary parts of the algorithm and what marketers or site owners should do to increase their rank, Google remains elusive on most of the 200 factors it uses to score pages and decide which page goes to the top of the results.
Even mathematicians familiar with the equations used to create the PageRank algorithm struggle with other non-numeric factors, Macworld reports. David Austin, a math professor at Grand Valley State University in Allendale, Mich., who published a paper on cracking Google’s algorithm, says the secret sauce is really a popularity contest wrapped in linear algebra.
“It’s like you’re having a popularity contest and you think everybody gets a vote, so I can vote for as many people as I want to,” Austin says. “So if I vote for 10 people, I give everyone 1/10th of a vote. So who wins that popularity contest?”
But then Google goes further. “They take a second pass through it, and look at who voted for who,” he explains. Google assigns a value to the importance of the site that casts the vote (or links to a site), and that site can pass on its popularity and importance to the site it linked to.
Gupta chisels away at the PageRank algorithm simply by looking at what the No. 1 ranked sites are doing. “We have identified 250 parameters that Google studies to rank a site,” he says. “We’ve got labs where people are constantly monitoring the impact of each. But the birds-eye view is, how can we make a site simply perform well in the natural course?”
The answer is, they don’t know, and Google ain’t tellin’.
Until somebody masters the Google algorithm it will grow as the most powerful–yet opaque–gatekeeper of information and commerce on the Web. And you know what they say about absolute power…
Google’s unofficial motto is “Don’t Be Evil,” but even if they were being evil–manipulating searches to quash competitors–how would we know?