How Amazon.com Reads Your Mind
| Category: Marketing Strategy | Date: 2001-03-12 |
When you buy a book from www.amazon.com, you are presented with a list of other books which "might" be of interest to you even though they are on completely different subjects than your purchase. And once you've set up an account, when you visit the website you will be greeted by name and offered another list of books amazon.com "thinks" you might want to know about.
It's an eerie feeling, especially because amazon.com is very often right. For example, when we selected Edwared Tufte's ornate "Visual Interpretations" amazon.com suggested a paperback called "The Design of Everyday Things" that we'd read about years ago, meant to buy, and promptly forgotten. What is going on here? Are the Luddites' worst fears of how computers will control our minds finally coming true?
As it turns out, amazon.com uses a process called "collaborative filtering" which was developed at the University of Minnesota in the early 1990s and is marketed by Steven Snyder, a Microsoft programmer who left that job to pursue a career as a psychotherapist.
Collaborative filtering presumes that buying habits can be a more accurate predictor of future behavior than other, more conventional research measures like age and race. If you enjoyed "Riverdance" I can bet you will also like "Bridges of Madison County" without knowing any more about you. Snyder's product, called GroupLens, automates this process on a mass basis.
Collaborative filtering isn't perfect, but it's successful enough to have gained a foothold at other sites such as www.cdnow.com, where shoppers type in the names of three favorite bands and will then be presented with a list of other titles and artists they are likely to enjoy. According to Chris Stevens, a senior analyst at Aberdeen Group who was quoted in the Wall Street Journal, "they have found a way to convert all this garbage that 99% of the world's population can't understand into tangible business value."
Snyder is also marketing GroupLens to call centers and retail stores, who might use collaborative filtering as a way to upsell buyers to additional products. As to the argument that his system might foster a new generation of consumers who all think alike, Snyder predicts exactly the opposite. "It suggests things you might like to experiment with...it's just the opposite of boxing you in." And that, indeed, has been our experience with collaborative filtering at www.amazon.com.
About the Author
DBMarkets@aol.com
http://www.msdbm.com
It's an eerie feeling, especially because amazon.com is very often right. For example, when we selected Edwared Tufte's ornate "Visual Interpretations" amazon.com suggested a paperback called "The Design of Everyday Things" that we'd read about years ago, meant to buy, and promptly forgotten. What is going on here? Are the Luddites' worst fears of how computers will control our minds finally coming true?
As it turns out, amazon.com uses a process called "collaborative filtering" which was developed at the University of Minnesota in the early 1990s and is marketed by Steven Snyder, a Microsoft programmer who left that job to pursue a career as a psychotherapist.
Collaborative filtering presumes that buying habits can be a more accurate predictor of future behavior than other, more conventional research measures like age and race. If you enjoyed "Riverdance" I can bet you will also like "Bridges of Madison County" without knowing any more about you. Snyder's product, called GroupLens, automates this process on a mass basis.
Collaborative filtering isn't perfect, but it's successful enough to have gained a foothold at other sites such as www.cdnow.com, where shoppers type in the names of three favorite bands and will then be presented with a list of other titles and artists they are likely to enjoy. According to Chris Stevens, a senior analyst at Aberdeen Group who was quoted in the Wall Street Journal, "they have found a way to convert all this garbage that 99% of the world's population can't understand into tangible business value."
Snyder is also marketing GroupLens to call centers and retail stores, who might use collaborative filtering as a way to upsell buyers to additional products. As to the argument that his system might foster a new generation of consumers who all think alike, Snyder predicts exactly the opposite. "It suggests things you might like to experiment with...it's just the opposite of boxing you in." And that, indeed, has been our experience with collaborative filtering at www.amazon.com.
About the Author
DBMarkets@aol.com
http://www.msdbm.com
Copyright © 2005-2006 Powered by Custom PHP Programming