We’ve all had a helpful bookseller recommend a book at some point. They have those freakish encyclopedia brains that remember every book they have on their shelves. The problem is, though, that more and more people are making purchasing decisions online and it is very difficult to replicate the helpful bookseller experience on the web. Online bookstores try with customer reviews and there’s the “people who bought x also bought y” style recommendations, but it doesn’t quite cut it.
Samantha blogged previously about Goodreads buying Discovereads.com to make recommendations based on your book preferences and those of your friends. Two other book recommendation engines have come across my radar recently as well: BookLamp and Whichbook. Neither is new, but they’re worthy of a blog post.
BookLamp
“BookLamp is a book recommendation system that uses the full text of a book to match it to other books based on scene-by-scene measurements of elements such as pacing, density, action, dialog, description, perspective, and genre, among others. In other words, BookLamp.org is a Pandora.com for books, based on an author’s writing style.” —BookLamp FAQ
BookLamp is a database of book content that can then be compared. So, you type in the title of a book you liked and BookLamp will analyse that text and find other similar texts to recommend. You can then vote results up or down and mark results as favourites.
The BookLamp algorithms look at what the team is calling Story DNA—the building blocks of the story that can then be compared to other stories. They further divide Story DNA into Story Setting and Story Actors. For more information on Story DNA, take a look at their post What is a Book Genome, really? What is Story DNA? And why should I care?
Answers the following type of question: “I really liked Harry Potter. Do you have something similar to that?”
My thoughts: An e-book retailer should be trying to buy BookLamp really quickly since those retailers already have a database of content and bibliographic data that’s just waiting for a recommendation algorithm like this.
Whichbook
“Every title on whichbook has been read by one of a changing team of 70 people who are drawn from libraries and literature organisations and come together to share training to create the entries. The ratings and comments are created by real readers who care about books.” —About Whichbook
Instead of starting with a book or author you already know you like as you do with BookLamp, Whichbook lets you start with ideas about the type of book you want to read. There is a list of 12 binary oppositions with sliders that represent ways you would describe books (like happy/sad, funny/serious, safe/disturbing, etc.). Or, you can browse based on character, plot, and setting elements.
Every book you find in Whichbook has been read by a real human and meticulously tagged. They also make it easy for you to find it in a library via the WorldCat database, or to purchase it via Amazon.
Answers the following type of question: “I’m looking for something funny and unusual. Oh, and it has to be short. What can you recommend?”
My thoughts: This is awesome, but probably not going to really take off because currently only 70 people are working on tagging books at one time. But I do think it could explode if the process was crowd sourced, letting anyone add or edit tags for a book.
Teaming Up
I wish these three distinct recommendation styles—preferences and social recommendations, comparisons, and descriptions—would join forces into one super-powered book recommendation engine.
What do you think? How do you currently find books online? And would you trust a machine to make recommendations for you?
