Customers Who Bought This Item Also Bought...

This morning I installed the latest update of iTunes, which includes the new Genius application. This feature allows you to choose a song from which Genius will instantly generate a playlist of similar songs from your music library, as well as make suggestions for additional purchases from the iTunes store. It's similar to Pandora, the Internet radio website which allows users to build radio stations based on particular artists. Unlike Pandora, however, Genius has the disadvantage that you can't give it feedback to improve future selections. A major drawback I've found to both services is that they can't distinguish more subjective traits like mood, thematic similarities in lyrics, time frame, or social contexts. For instance, both services totally miss the mark when it comes to The Smiths. There's a whole genealogy of bands heavily influenced by or downright imitative of The Smiths' moodiness, bookishness, queer sensibility, etc. that get passed over in favor of songs that Genius perceives to have the right rhythm or combination of instruments.

This is a (maybe digressive) way of coming to my point that whatever their faults, these sorts of recommendation tools are becoming more prevalent in my life, and I wonder what role they are going to have in libraries in the near future. I don't think it's presumptuous to guess that they will have some role, considering the countless other ways that chain bookstores have shaped the public's expectations of library services. Both the Barnes & Noble website and amazon.com make use of a recommendation feature, and although I don't work in a library, I'd wager that many librarians are already using these sites in their everyday work. I know that I used them all the time as a bookseller, ironically at a fiercely independent academic bookstore. Despite subscribing to multiple other databases like Books In Print and BookSource, we all had to admit that amazon.com was easier to use and provided multiple features (like cover images, tables of contents and reviews) that the others lacked.

And yes, when we found ourselves having to make recommendations beyond our personal areas of expertise (or more snobbishly put, below our areas of expertise), we would sometimes fall back on those automatic recommendation tools on amazon. While I don't think it's a bad thing for booksellers and librarians to use these tools, I do think it's important to have a sense of perspective about their advantages and disadvantages.

First of all, amazon.com is going to make every attempt to sell me a book, not caring at all that its recommendations may become redundant or illogical. For instance, I once bought a copy of Proust's In the Shadow of Young Girls in Flower on amazon.com. This was maybe four years ago. To this day amazon recommends books based upon this title including the very same title in a large print edition, an audiobook, a Canadian edition, and a hardcover edition. Why I'd want to buy the same title in five formats is beyond me, but it makes sense for amazon.com to give it a shot. Swann's Way is also heavily in the rotation of recommendations, which they might have logically deduced I don't need since it's the volume in the series that comes before the once I bought; I'm told I may also like Within a Budding Grove, which is the book I already bought in a different translation. There can be a lot of weeding to do to get past the repetition.

There are also places that amazon.com excels at, however. I recall that amazon.com could come in really handy with book groups. People would tell me they'd just finished The Kite Runner and want to know what to read next. I could query Amazon and find that customers who bought The Kite Runner also bought Water for Elephants, Reading Lolita in Tehran, and The Mermaid Chair. Sure enough, those were popular choices with the book groups in Amherst as well. While it did feel a little lazy to me to do this, and while I felt guilty for making generic and popular suggestions, in truth I knew that this was often what the customer wanted. Asking for a book recommendation was often overtly synonymous with asking what everyone else was reading and what sold well. In this sense, amazon.com was better equipped to make recommendations to this cohort of readers than I was.

Of course, those weren't the recommendations that I was proud of making. I much preferred the occasions when I knew from experience just the exact shade or mood or sensibility that attracted a reader to a particular book, and actually felt a sense of investment in what she should read next. But much of the time it was all I could do to recommend on the basis of a similar tempo.