Many platforms are chasing the recommendation Holy Grail, but a new company called Small Demons, which is scheduled to launch at TOC Frankfurt, is taking a different approach. Instead of providing recommendations, Small Demons uses metadata from within the texts of the stories to facilitate discovery — not just of other books, but movies, music and other types of pop culture woven into the books. In this interview, Valla Vakili (@Sideposts) founder and CEO of Small Demons and a speaker at TOC Frankfurt, talks about how the platform works and how it can be useful for publishers as well as for readers.
Our interview follows.
What does Small Demons do and how does it work?
Valla Vakili: Discovery is the ultimate problem we’re trying to solve and the ultimate value we’re trying to create. If you’re a reader and you go online to find a book related to the book you’re reading, you have some very standard options: the “people who bought this also bought that” function from Amazon; ratings and reviews from people you may or may not know. I feel like a lot of this is pretty washed up. And the thing that’s being substituted for it now is the social signal — your friends are reading this; your friends are reading that. That’s pretty weak, too. There’s a much stronger source of discovery, which is the book you’re already involved in.
Here’s a personal example from another medium: I recently saw the Woody Allen movie “Midnight in Paris.” In it, there’s a brief exchange with a character playing the Spanish filmmaker Buñuel. I’ve had one of his films, “The Discreet Charm of the Bourgeoisie,” sitting in my Netflix instant queue for a long time. If a friend of mine said, “Oh man, I watched a Buñuel pic over the weekend, you have to watch it,” I probably wouldn’t because I’m burned out on recommendations from friends. When Netflix recommends it to me, I just put it in my queue and I don’t watch it. But when I’m in the context of “Midnight in Paris” and I’m caught in the grip of that story and the Buñuel character catches my attention, now I want to go watch something by him. It’s context, and I’m hooked. It wasn’t even the movie mentioned in "Midnight in Paris," but I’m hooked.
This sort of contextual discovery is much more powerful because, by nature, it’s not siloed. It can open you up to things in different mediums; it’s an extension of the story you’re already in. For years, I found myself doing this in books. I would mark up books when they had references to interesting music, interesting places to go, interesting movies, interesting drinks. We thought if we could bring this all together and connect stories to the things they talk about, then it would open up this new form of discovery where you can continue down the path of the story that you’re already in and find the things that give it the texture you liked in the first place. So, for example, if you have a book like “High Fidelity,” which has a lot of music in it, we give you the added experience of discovering the music — as well as all the other pop culture woven so deeply into it.
So, it’s discovery via the text of the book itself?
Valla Vakili: Yes. We go through and create a graph for all of the little things inside of the books — the things that lead you off to new places — and then we show you all of the books that share those same elements. Once you’ve read the book, you can decide that you just want to go get the music, or you can decide to go get the music and then discover other books that have similar kinds of music in them. It’s two types of discovery: The first takes you deeper into the world of the thing you’re already in — places and things and such — and the second leads you toward books like the one you’re reading based on the objects that we’re graphing.
Can readers access the Small Demons context “live” while reading an ebook?
Valla Vakili: Not now, but down the line. That’s one of the use cases for our API. I think there’s plenty going on in terms of changing how people read already — adding things to books, making the digital book look like something other than the print. To me, the stories are already great, and they have this incredible ability to lead you out into other things. For us, it’s not about shoving more inside; it’s about using the story as a way to find all of these other things. So, this is a use case for us, but it’s a later use case.
How are you gathering your data and mining the content?
Valla Vakili: We have publisher relationships, and we mine EPUBs. We have kinds of topics that we’re looking for, and we identify those topics, add detail to them, map them back against our taxonomy, and then it’s an ever-continuing process of refining that. But that’s the basic process — we are looking for certain entities that, broadly speaking, fall into the people, places and things categories. We started off doing this in a manual fashion to determine what exactly we were looking for and where the cutoff was between what we wanted to express and what we wanted to let go. We’ve moved away from that now to a machine-based approach where manual curation is used for verification and refinement but extraction is machine based.
Will you share datasets with publishers?
Valla Vakili: Yes, we will be sharing our datasets with publishers. Discoverability and the need for richer metadata have been big topics in our discussions with publishers. What we’re providing is a very different type of metadata. It’s from the substance of the story itself, and it expands the range of available information about any given book and the connections between other books that share that same information.
What about BookLamp’s recommendation mechanism — is it similar to what you’re doing, or does Small Demons have a different approach?
Valla Vakili: I think it’s a completely different approach. We are actually not recommending anything, if that makes sense. We’re not saying, “If you like this book, you will like these other books and things.” What we’re saying is, “This book you like, it has all these things in it. They’re also in these other books. Enjoy. See where it takes you.” And then we let the readers decide what it was that attracted them to the book and what they may want to go experience next — it could even lead them to something other than a book; it could lead to the food, the music, the movies that are somehow related to the book.
What’s your launch schedule?
Valla Vakili: We are launching at TOC Frankfurt. The first thing that we launch will be the website. There will be other experiences and platforms that come later.
This interview was edited and condensed.