Wired’s How-To Wiki offers some pretty good strategies for improving the recommendations you get from Amazon. Here’s one of five tips:
Rate items you like. Sure, it feels good to trash a book you hated, but telling Amazon what you enjoyed will generate better recommendations - and the more, the better. On hold with the cable company? Rate a few items. On a conference call? Rate a few more.
It’s all good advice, even if you don’t use Amazon, because nearly all e-stores use similar strategies for making recommendations.
That said, once you’ve made a lot of purchases at Amazon or any other store, you almost never discover new ideas through the recommendation engine because the only things on your list are more books/music by authors/artists you’ve already bought.
This strikes me as absurd. The whole point of making recommendations is to help me discover things I wouldn’t have discovered on my own.
I don’t need a complex computer program to tell me that if I enjoyed one U2 album, I might like another. I can do that on my own.
I need a computer program to tell me that if I bought an album by U2, I might like to read “Wide Sargasso Sea.” (Or whatever — I’m just making up a distinctly non-obvious relationship.)
I thus call upon Amazon and its fellow e-stores to abandon the low-hanging fruit of same-author picks. It would certainly add more value to my dealings with them.
Popularity: 3% [?]
Entries (RSS)