It’s easy to forget that Oyster, dubbed “The Netflix for Books” by every writer who’s covered them, has only really been around for a year. In that time, they’ve signed on two of the Big Five publishers, built a library of 500,000 books and basically made us forget that Scribd has been trying to contend for exactly the same space.
And at the heart of Oyster’s business is their recommendation engine — a recipe for bringing readers in and hooking them on book after book. On one end of that engine is their editorial offerings — curated collections that go beyond genre and into moods or themes like “The Science of Everyday Life.” This morning, they added more “book list” features, essentially allowing everyone from users to major authors to make their own literary playlists and share them with followers and friends.
On the other end is a data engine that can examine your reading habits, determine your pace and preferred genres, and build a profile specific enough that it can know just by what phone you have if you’re more likely to like The Hunger Games or How to Win Friends and Influence People.
We spoke to Oyster cofounders Eric Stromberg and Willem Van Lancker at their NYC headquarters about how they’re using reader data to build the new book subscription incumbent, and how they’re protecting themselves from the titans who are coming for their throne:
* * *
So what data from readers are actually useful for you?
Eric Stromberg: We look at how far through a book they’ll read, what books are opened, how fast they’re read, and then do cluster analysis around different readers. For example, romance readers are twice as likely to finish reading a book front-to-back. We look at what percent of readers finish 10 percent, 20 percent, 30 percent, all the way through.
Willem Van Lancker: There are definitely some interesting trends in books. So everyone might start this one book, but no one finishes it. There are books that people just flip through really quickly and don’t stop, books that are read in longer sections, that kind of thing.
So when you find these clusters, who gets to tap into that data? Is any of that making its way back to publishers or the authors?
WVL: That’s something that’s definitely on the agenda. To date, it’s been a total black box on the author’s side in terms of who’s reading the books, where they are, when they’re sold, where people drop off. They’ve never had access to that any of that type of data.
Shouldn’t there be a concern for that kind of data influencing editorial?
ES: We’re so far from that. Right now, authors and publishers don’t know how those books were read, and sometimes they don’t even know how many were sold. To start to unlock a little of that information is interesting, but to suppose that data will change an author’s creative process anytime soon is a bit presumptuous.
But you’ve already built that infrastructure. If you can really look at a book and see what the completion rate is, why can’t an editor going in and come up with three titles and A/B test them like Upworthy headlines?
WVL: Actually, of the more low-hanging examples is cover design, because some books already have ebook covers and print covers. But our notion of success isn’t Upworthy’s — it’s not about watching a two-minute video, it’s reading a five to eight hour book. The book’s gotta be good.
ES: Some authors might want to be informed by the data, but they won’t have it impact their writing style. I’ve already met many authors who spend half of their day writing and they spend the other half of their day marketing their books on social media, sending emails to their fanbase, whatever.
Are you working with authors yet on exclusive content? Netflix, Amazon Prime, Kindle and Spotify all have shops for making original work to draw people onto the platform.
ES: It’s not on the near horizon to do original content of any sort. Netflix has started doing that, but it’s decades into their business. Publishers produce great content, and what we’re really good at is distributing that content and making a product to help people read and discover books.
Speaking of Kinde, Amazon recently tried edging in on subscriptions.
ES: That’s not a surprise for us, and Amazon really shined a light on the market. They’re making people aware of unlimited subscription, which is a fundamentally better experience on Oyster.
WVL: It helps for us to be compared to something that’s not as good.
But what about when Apple enters the space? They bought three recommendation engines in a two week span. Since the Booklamp acquisition, it’s obvious they’re headed there.
ES: Well, personalized recommendations are valuable both in a subscription world and an à la carte world. We’ve found success with this model, but we can’t speculate about what Apple is going to do.
Still, look at everyone working the shadow of Uber: someone will start the “Uber for jets,” Uber will show up like, “Psych! Uber is the Uber for jets and you’re out of business now.” Isn’t there some peril when someone like Apple can replicate your entire business overnight, only with more infrastructure and experience in publishing?
WVL: There hasn’t been an à la carte retail player that’s pivoted into unlimited subscription easily. You could say Apple, maybe, but they had to buy a company to do that. Amazon certainly has proven that they’re not able to do that both in music and movies, and the startups that innovated in those spaces are leading there.
ES: This happened when Facebook started taking off: There would be a Facebook for sports, but then it turned out that Facebook was just the Facebook for sports.
There are already smaller startups edging into your space, too, like Epic!, which is like the Oyster for kids books. Early stage startups always pretend there’s enough room for everyone in their space, but that just ends up not being true.
ES: I think it’s yet to be seen. If you look at music, Pandora is a big public company, Spotify is on its way to being there and iTunes is still a huge business. And the music market is slightly smaller than the books market.
WVL: Also, there are different kinds of readers. There are people who want to learn programming, there are people who want to use it as an escape mechanism, purely for entertainment, people who use it for other forms of education, to read with their children.
ES: We’ll certainly focus our intention in and zero in on more of those genres over time.
WVL: But it also might be the Oyster for Kids books is Oyster.