While bestseller lists, review columns, and celebrity-endorsements (ahem, Oprah’s Book Club) have been the standard for finding new books to delve into, Aaron Stanton and the folks at Novel Projects are trying to shake things up. Using a data-driven approach, Novel Projects’ website, BookLamp, will offer you book recommendations after algorithmically analyzing your reading preferences. Examining factors such as pacing, language, and even plot setting, BookLamp may know your reading taste better than you do. And because the website works entirely on quantitative data, you’re just as likely to stumble upon a book from an up-and-comer author as you are from Stephen King. To celebrate BookLamp’s official launch next Tuesday, I sat down with CEO and founder Aaron Stanton.
Besides BookLamp, are there any other projects that Novel Projects Inc. is working on?
There are actually three different branches of what we do: what we call the Writer’s Project, the Reader’s Project, and the Publisher’s Project. The readers project is the Pandora-for-books model. The Writer’s Project is what individuals use. What we do is say, “let us analyze your manuscript,” and we will tell you, “here’s five editors who historically have published books like yours.”
And the Publisher’s Project?
What’s funded the company for the last several years has been what we call the Publisher’s Project, which is almost a consulting model. We have all this information about the book so we can compare it to other books that have come before it and help [publishers] improve where to target their marketing.
How many different qualifiers does BookLamp use to analyze and recommend books?
We continuously build it. But at our re-launch – our revamp of the site – the number is 132 thematic ingredients.
Is that how you refer to the different components – ingredients?
It’s a poor analogy when applied to art, but one that we use occasionally because people seem to understand it. I describe it this way: If you had a chocolate cake that you really liked, and you wanted to find another chocolate cake very much like it, you’d want to know a few things. You’d want to know what the ingredients are. And you’d also want to know how much of each one so you don’t get too much salt and too little sugar. We are consistent in saying that we don’t judge books. I would abhor the idea of trying to judge quality using a computer. All we do is, if you come to us and say, “I like this book,” we’ll try to find you another book similar to it.
Besides the 132 thematic ingredients, you said you also look at other aspects, like the book’s language?
What we call the core elements of language is what we break down into Pacing, Density, Description, Dialogue, and Motion. And then we have a whole slew of what we consider non-core, like Perspective. So what we do is take the core elements, and based on what books you’ve told us you like, we try to figure out for you what your writing style preference is, and then say, here’s a bunch of recommendations. In total, our system stores more than 32,000 points of genetic data for every book that we analyze.
With the site’s launch, is BookLamp going to be on a subscription basis?
No. We are not aiming to be a commercial site. The goal of this website, more than anything else, is that we want to recruit publishers. This is a cool project. We think it has a lot of potential. No matter what we have, it is dependent on how many books we have in the database.
This is a big demonstration of what we look at when we see books, and the way we think this data can be used by a reader to discover new things that aren’t so heavily driven by marketing dollars. So a new author with a new book is as equally weighted as Stephen King, based entirely upon content inside the book. This idea that what’s between the covers is really what matters is one that I think is obvious, but I think it is overlooked a lot of times.
When the site launches on Tuesday (August 16th, 2011), how many books will be in the database?
Is there an interesting back-story as to how you got the idea for BookLamp?
When I was 16 years old, I wanted to be an author. I’d write short stories and I’d force my dad to edit them for me – poor guy. And at the top of every single page, on a scale of 1 to 10, I’d have him write the “Interestingness” of the page.
On each page? Not each story?
Each page. And then I’d graph it. The idea was to try to get a story with as high an interestingness on every single page. So when I got to college I was trying to figure out ways of using the computer to do that – to compare my writing to people I considered my influences, like Stephen King or Michael Crichton. What I wanted to know was… what is the average length of the first scene in a best-selling book, compared to ones that have not been best-selling? Is there a difference? I wanted to know what the rise and fall of the Pacing and Density was.
You mentioned a couple of authors that are influential to you. Any other favorites?
Orson Scott Card. When I grew up, Ender’s Game was fabulous. There is a whole series of authors that I… yeah, there’s a long list.
When I log on to Amazon and browse books, I’ll get suggestions. What do you think is flawed with that system, or what’s the problem there?
I’m not sure I’d call them flawed, just with different strengths. One of the biggest concerns with social recommendation engines are their natural pyramid structure. Marketing dollars drive awareness; awareness drives what’s recommended and discoverable. There’s a huge quantity of books on the bottom half of that pyramid that people have a hard time discovering.
Do you think this project would have been any different – for the better, or for the worse – had you not done Can Google Hear Me?
Oh, it would be tremendously different, and for the worse. The response we got from Can Google Hear Me… we got 4,000 “good luck” emails in a two-week period. At that point, you can’t stop. There are 4,000 people that wished you luck on a project. You can’t email them back and be like, “I gave up. Sorry.”
Born and raised in Idaho?
No. Actually, I was born in… either Ames, Iowa, or Greensboro, North Carolina. One of those I’m born in, and one of those my brother is. I think I was born in Ames.
Seriously? That’s pretty funny.
That’s probably the more forgivable one. I’d have to think a little if you asked me how old I am. I think I’m 29. Born in 1981. I might be 30.
Did you not have your birthday yet this year?
Um, you dramatically underestimate how busy a startup is. But, yes, I had a birthday in June, so I’d be 30.
So how can readers who stumble upon this article contribute to the success of BookLamp or Novel Projects, Inc.?
Giving us feedback, telling us how we are doing, is a tremendous part of it. What I’d really ask is for their help in getting our message out to the right people in the publishing industry. If a reader knows someone who knows someone that might be the right ears for us in a publishing house, fire them an e-mail and say, “Hey, look at these crazy guys…” That would be tremendously helpful for us.
Anything else you’d like to share?
I will throw in this: that while I spend a lot of my time doing these kinds of interviews, none of what we do would be possible without some extraordinary talent and skill of the people who are working day-to-day around me.