If you suffer from a Millennial attention span then I suggest — as if the suggestion were necessary — you stop reading now. Because this is a post about books. Thousands and thousands of books.
Valla Vakili loves books. Specifically, he loves the connections between books: the places, people and things that crop up across multiple titles. The connective tissue that binds the galaxies of fiction and non-fiction into one vast shared universe of stuff that’s mentioned in books. He calls this parallel universe “the Storyverse”
Take, for example, Las Vegas. The book that instantly springs to mind, of course, is “Fear and Loathing…” But the city has been featured in hundreds of titles from Don DeLillo’s “Underworld” and David Foster Wallace’s “Infinite Jest” to Chuck Klosterman’s “Sex, Drugs, And Cocoa Puffs” and David Carr’s “The Night of the Gun”. Interesting, huh?
Valla Vakili thinks so. Likewise he finds it interesting that Cocoa Puffs connected Klosterman to Jodi Picoult’s “Handle With Care”.
“I didn’t care when you ate all the Cocoa Puffs so that I had to have Frosted Mini-Wheats as a snack after school. All that mattered was that at 4:30 p.m. I was getting my braces off, after thirty-four months, two weeks, and six days.”
See? (And don’t get me started on the role Frosted Mini-Wheats play in Stephen King’s “Duma Key”.)
And so the former VP of product at Yahoo joined forces with former Director Product Design at Yahoo Tony Amiedei, former VP Data at MySpace Hala Al-Adwan, and MySpace Chief Data Architect Christa Stelzmuller to figure out out a way to analyze every book in print and track the millions of possible connections between them.
The result is Small Demons.
And there goes your afternoon.
I was introduced to Vakili by my friend Richard Nash, formerly of Soft Skull Press and now Small Demons’ Director of Content and Community. Nash’s recommendation carries a lot of weight: Too many startups that promise to revolutionize, disrupt, or otherwise fuck with literature are run by people who one suspects have never actually read a book. With Small Demons one imagines the founders must have struggled to drag themselves away from their libraries long enough to actually build a site and create a company.
But create a company they have. Small Demons raised an angel round in 2010 and a seed round in 2011 from investors including Yuri Milner’s Start Fund, The Social+Capital Partnership, Camp Ventures, XG-Ventures, Craig Shapiro, Greylock EIR Alison Rosenthal, ex-Facebookers Tim Kendall and Julia Popowitz, and ex-Googler Harry Cheung.
Just as importantly, the company has partnered with more than a dozen (and counting) publishers including half of the “Big Six,” Simon & Schuster, Random House, and Hachette. (Only Penguin, Macmillan, and Harper Collins to go.)
The deals give Small Demons access to the raw digital files of those publishers’ entire catalogues — and in return gives the publishers access to all of the resulting data generated by Small Demons. In an industry that remains fixedly protective of its digital intellectual property, it’s hard to overstate how impressive these deals are.
I said as much to Vakili when we spoke a couple of weeks back. The company was preparing for the London Book Fair, where they were promising to announce yet another partnership (which turned out to be with Random House UK). I also asked him to explain some of the technological challenges involved in connecting the entire world of books.
After reminding me that he isn’t the company’s CTO (that would be Al-Adwan), Vakili nonetheless gave a great description of how everything works. In a nut: the digital files are ingested into Small Demon’s proprietary technology and scanned for known places, people, things, brands, and anything else that might be a possible keyword. Then the keywords are disambiguated to distinguish between, say, Paris the place, Paris the Hilton, and Paris the would-be husband to Juliet Capulet.
“Paris is what we call a ‘known offender,'” says Vakili. The company’s algorithms have got good at disambiguation, he insists, using a combination of existing Small Demons data, external data sources and even open-source encyclopedias like Wikipedia. But still some words are flagged for human intervention — just to be sure.
The company has processed 3,000 titles so far and is currently “working through several thousand more,” but there’s still a long way to go. I was hugely offended to discover, for example, that my latest book isn’t yet included, despite mentioning Las Vegas — though to be fair it makes no mention of Cocoa Puffs or Frosted Mini-Wheats.
Processing millions of books will take time, but also money. I asked Vakili if, investment notwithstanding, there was a real business behind “the Storyverse,” or whether book lovers should be making the most of it while they can. Right now advertising is a non-starter: Small Demons only recently moved out of invitation-only beta with a low five-figure userbase. By contrast, book review site Goodreads has six million registered users. Even running at full whack, Small Demons is never going to be a huge ad-supported business: It simply isn’t dumb enough.
So what is the revenue plan? Vakili gave me a choice: Either the standard “we’re not ready to announce that” answer on the record, or the full story off the record. I opted for the latter, providing Pando gets to be the first to announce when they’re ready. Deal.
So here’s what I can share: There’s a real business model here, but it’s far from guaranteed. Some of the ways to make money from a book site are obvious, but Vakili and his team have figured out a second, third, and maybe even fourth layer of monitization for their “Storyverse”. The fourth level, I hope he won’t object to me telling you, involves tapping a massive seam of revenue that publishers simply can’t (I hope) even consider. We. Shall. See.
In the meantime, though, the service remains totally free and unencumbered by commercial exhortations. If you love books and have a spare — I dunno — lifetime, you’ll probably want to take a look.