We spoke to Ali Albazaz, CEO and founder of Inkitt, a data-driven publisher that uses an artificially intelligent algorithm to analyse online users’ reading patterns. With a background in computer science, Ali believes predictive analysis has the potential to make the publishing industry more fair and objective. We asked him about how he came to form his company with designer Linda Gavin, the value of data in book agenting and how they built up the site’s core audience and readership.
Inkitt is a ‘data-driven publisher’ that curates writing submissions for online publication using an artificially intelligent algorithm to analyse reading patterns. How did the idea for Inkitt come about? Tell us a story from your early days.
I founded Inkitt together with Linda Gavin and it was launched to the public back in February 2015. Both Linda and I love reading fiction and discovering new authors, and we kept having those conversations about how, even though the publishing industry is constantly evolving to keep up with how people choose and buy books, the decision-making process in terms of which works will get to see the light has remained the same for centuries. The industry relies on the experience of editors and literary agents, their knowledge of past market behaviour and their gut feelings to decide whether a novel has the potential to become the next bestseller. This has resulted in initially rejecting works that ended up becoming huge successes: for example, you might have heard that J.K. Rowling’s Harry Potter and the Philosopher’s Stone was rejected 12 times, Stephenie Meyer’s Twilight 14 and Stephen King’s Carrie 30 times before they made it. Which got us thinking: shouldn’t there be another, fairer way to make decisions? A process that takes into account what the audience actually thinks instead of making assumptions on behalf of readers?
This is how Inkitt was born, with the vision to become the world’s fairest book publishing house and democratise publishing by putting the decision in readers’ hands. Our mission is to help aspiring writers find an audience for their work and kick-start their career. The Inkitt algorithm can predict future bestsellers by analysing reading behaviour: if readers love it, we publish it.
Inkitt sometimes acts as a stepping stone for writers to get picked up by traditional publishing houses. What does Inkitt add to author advocacy that might have been missing in an industry solely reliant on agents and editors?
As I mentioned above, Inkitt is all about reader-driven publishing. We help great works by new and talented authors get found and published by relying on reader’s opinions instead of middlemen and the (oftentimes subjective) view of a single expert. The data is out there, so why not gather and analyse it to make a decision, instead of relying on assumptions, personal preferences and hunches? Harry Potter was initially rejected by 13 publishers who felt there would be no audience for it out there. Instead of making assumptions, why not observe readers as they’re reading a novel to understand whether they like it or not? This is exactly what our algorithm does.
The Inkitt algorithm analyses over 1,200 reading behaviour dimensions to understand whether readers like a novel or not. Based on that analysis, we identify top performing novels and offer them a publishing deal. We first publish a novel in eBook format and, if it performs well, we proceed by either publishing the print book in-house or pitch it to A-list publishing houses. For example, we’ve teamed up with Tor Books to publish Erin Swan’s YA novel Bright Star, which is coming out next summer.
To cut a long story short: when Inkitt pitches a book to a publishing house, we already have dependable data that proves how strong a potential for success a new novel has. As a result, the novel becomes a lot more attractive to a publisher as we’re taking the guesswork out: we already know that readers love it, so it clearly is a win-win situation. What is also great for the author is that we get better terms negotiated.
Do your authors have access to the data collected on their books for future iterations? How does Inkitt work with the authors they select?
When it comes to providing access to collected data, this is a feature we’re thinking about a lot. We want to make the stats easy for authors to understand and offer actionable insights. Right now, we’re working on a few dashboard ideas for this.
We also provide tips and guidance on how writers can spread the word on their work and get more readers. For example, we offer them a checklist listing Facebook Groups on their dashboard, where they can share a link to their story and request feedback. In addition to that, we have several writing groups on Inkitt where we invite published authors to share their experience and answer writers’ questions on writing, getting feedback and getting published.
If our algorithm picks a novel as a future bestseller, then we offer the author a publishing deal which includes editing, cover design, if desired, and a dedicated book marketing team with an initial promo budget of $6,000. The writer receives 50% royalties. If their book doesn’t sell at least 1,000 copies in 12 months and they’re not happy, they can cancel the contract and get all their rights back. It’s worth mentioning here that the first book we published sold more than 1,000 copies just on the day it launched! Finally, in the case where we don’t publish a book in-house but partner with a publishing house, the author will receive 85% of the net revenue.
In addition to internally being a place for data collection, the site is popular amongst readers keen to find good writing in styles and genres they enjoy, for free. How did you initially build up these recurring visitors and readerships?
That’s very true. Inkitt also has an engaged fiction readers’ community that counts over 700,000 members. And we’ve managed to get there in just 18 months from launch!
So how did we do it? It mainly goes back to the fact that I initially built the platform together with 100 authors who shared my vision and were there for Inkitt before there even was a product. They were all super supportive and their help with spreading the word about Inkitt was invaluable.
Another thing that was equally important was the fact that engaging with the Inkitt community has been one of our top priorities since day one. Back when there were only four of us in the team, Linda spent a large proportion of her time responding to user queries, discussing issues with them, understanding what they love about Inkitt and what could be improved and asking for their thoughts. Right now there are 20 of us and we have a dedicated team who are always there for our readers and writers and deal with any issues that might arise. Feedback is really important for us: we regularly ask our members what their thoughts are, keep track of their suggestions and evaluate and implement them as we move forward so we can always keep improving the experience we offer.
Last, but not least, it’s about knowing who your core audience is. When it comes to readers, we wanted to become the go-to digital library for fiction lovers who want to discover great works by emerging authors and be among the first to read them before they go mainstream. Inkitt is the ‘Hipster’s Digital Library’, and we achieved this by focusing on that specific audience, listening to their needs and desires and responding to them.
Your company was clearly founded on a good deal of technical knowledge and research. What are your thoughts on similar endeavours like the recent book The Bestseller Code (profiled here)?
Indeed, the Inkitt algorithm is at the core of our mission to democratise publishing. I’ve studied Computer Science and started coding at the age of ten. I wrote the first lines of the Inkitt algorithm and now have a team of four experts in Development, Machine Learning and Data Science who work on further optimising the Inkitt algorithm and our data analysis process.
In regards to The Bestseller Code, we actually held an AMA (Ask Me Anything) with the authors, Jodie and Matthew, on Inkitt recently. I was really interested in their work since it’s about identifying patterns to understand audience preferences: by taking into account which elements seem to affect how well a book is received by the audience, writers might be able to better ‘design’ their novel. So it’s again about harnessing the power of data, although from a different angle compared to what we do, since at Inkitt we analyse reading behaviours whereas The Bestseller Code is about analysing the text itself.
I’m now halfway through the book but I’d prefer not to make a statement on it before I’ve actually finished it.
What are your 3 favourite books of all time?
The Alchemist, by Paulo Coelho
Caged, by Onaiza Khan
Bright Star, by Erin Swan