Statistics as a business process

I'm not really an unbiased observor in this field but smart companies are betting their companies on statistics rather than intuition, and its paying back in greater dividends for their investors.

Examples include:

Amazon: The space for tabs on the home page is highly limited. Thus, department heads at Amazon are under intense pressure to get the tabs they are responsible onto the main page. In a normal work environment this might result in creative and inefficient political means to determine which tabs get on. Almost no incentive system can solve this without being gameable. And apparently that's what happened until someone got the bright idea of using simple bayesian statistics to determine which tabs optimized the companies top-line revenue the most -- these tabs then get the precious space. Get that? The company does live A - B testing and automatically slurps up the results and without interference picks the profit maximizing tabs.

More later...

Ken Kalfus

One of the advantages of going to fake graduate school is that I get a fair amount of time to read again. Especially so given the interesting status of the local mass transit solutions. The latest book I read is by Ken Kalfus. Ken Kalfus doesn't live in Brooklyn. Or even Berkeley. Ken Kalfus has lived in Paris, Moscow, Dublin, Belgrade and Philadelphia. But not Brooklyn. Which is the only reason I can think of for why you haven't heard of him. Which is a damn shame because you're missing something special.

Ken writes with a great capacity for understatement that's really quite refreshing in the day and age of David Foster Wallace (who is - no doubt - quite a treat himself). Ken's two anthologies Thirst and The Commissariat of Enlightenmentare some of hte best short stories I've read in ages, his new novel A Disorder Peculiar to the Country isn't quite up to these standards. But its still a wonderful piece of post-9/11 literature, almost as timely and accurate as Safran Foer's Extremely Loud and Incredibly Close.

There are plenty of synopses around, so I won't burden you with that. DPTTC, like EL&IC, focuses on the irrationality of New York City, and by proxy, America during the year following the attacks on the World Trade Center. Characters respond to the incomprehensible loss of life with a numbness to the world around them, mindlessly hurting those around them with their narcissism and need for comfort. They also both are over-engineered around fairly obviously metaphors and symbolism but make up for it by daring to tap into emotional responses from readers. Where DPTTC breaks down is in the pacing. Kalfus will spend pages and pages on a a lovely party scene that breaks one of the characters out of his stupor just long enough to care about the feelings of another woman only to then resolve the entire book in just a few pages, eliminating all the emotional resonance from the story just as we've come to really understand how sad and forlorn our protagonists really are.

Read it. But read EL&IC first.

What's your score

My score is 16 / Or I'm only 4.5 times too inefficient for the world to be populated by people all like me. A great write up on what these numbers mean can be found here.