Blog Home
Written By
January 21, 2011

Amazon's Algorithm: Am I what I buy?

January 21, 2011

I racked up quite a few purchases before Christmas, and now I’m facing the New Year hangover. The kind of hangover I’m describing is not a headache from the champagne but the feedback noise of recommendations for products.  These recommendations are generated by the secret software that Amazon and other companies use to analyze what I’m buying and predict my tastes. SEO's, SEM's, Social Media gurus and PPC managers are all in this game to a certain degree, but Amazon--because of their captive 'audience' or clientele--has perfected it to an art.

It can be a little shocking, sometimes, to realize what kind of impression I’ve made on Amazon. Who does Amazon think I am? Clicking through the model of my tastes that they reflect back at me can be like looking into a funhouse mirror.  Some of their choices seem oversimplified or even patronizing.  Some seem random.  But then some of their predictions are dead on.  It can be irritating to be sloppily profiled by a machine.  On the other hand, it can be unnerving to be accurately profiled in subtly nuanced and minute detail.  These are the vicissitudes of living in symbiosis with Google-bots and Amazon algorithms.

All the books that are being hawked to me lately are on the oversimplified side of Amazon’s opinion of me. Because I browsed the Amazon listings for The Road, the Cormac McCarthy novel that inspired the 2009 movie-adaptation, as well as Neil Gaiman’s eschatological comedy Good Omens, I’m getting all kinds of weird recommendations.  Due to the fact that both books have a post-Apocalyptic feel to them, the information crawlers inside the Web classed me as some sort of  end-times nut. When I log on to my account, books from the Left Behind series pop up.  I don’t own any of the Left Behind books. Nor do I have any desire to purchase them or any other book by Timothy LaHaye. Especially not now that my computer thinks I might want to.

But then, oftentimes, the algorithms peg our tastes with Vulcan-minded precision, humbling us with their ability to transform our beloved individuality into a set of mathematical near certainties. A few weeks ago a friend of mine logged onto his Amazon account and saw a recommendation for an album that his wife, a singer-songwriter, had just released on a label that he used to be employed by.  The dead-on, digitally determined pick was almost telepathic, as are, to my mind, many of the song selections made for me by Pandora, the online ‘radio-equivalent’ that plays on my Mac while I sit at my desk. One minute a melody will float up into my head, and the next moment it will float out through my headphones. Maybe I’m an android after all. Maybe the memories and desires stored on the internal hard-drive in my head are easily hacked.  For good or for bad, the algorithms and I have gotten to know each other pretty well over the years.

Related Posts