If you‘d have once predicted that a piece of fan fiction inspired by tween vampires and laced with sadomasochism would become one of the bestselling books of all time, your credibility — and sanity — may well have been called into question.
But then, there’s Fifty Shades of Grey.
While author E.L. James isn’t exactly known for her technical skill or style, her stories, it turns out, consist of certain key ingredients that often lead to success in the erotic romance genre: a smoking hot, albeit complicated, rich man; a plain brunette he’s infatuated with; and a wild love tango that ensues.
There are even books dedicated to this concept: The Bestseller Code, written by Matthew Jockers and Jodie Archer, explores these types of fiction formulas, positing that many bestsellers share a particular DNA structure — and using machine learning algorithms to make its case.
Now, Austin startup AUTHORS.me has built a similar data mining tool called StoryFit, which is designed to predict how well an audience will receive a book before it's even published.
“StoryFit is AI technology that analyzes book manuscripts and movie scripts to identify hits for publishers and studios,” said co-founder and CEO Monica Landers. “We identify what’s going to work, why it’s going to work, and how best to market the property.”
The platform ingests original content and gives insights to commercial viability, tone, style, story arc, pacing and marketing recommendations using keywords and metadata.
They’ve even recruited the help of data scientists Mark Bessen and Grace Lin to help lead the initiative.
The duo makes a compelling pair. Bessen, who joined the team from Apple’s iBooks, actually worked with Jockers and Archer on The Bestseller Code, and Lin, who has a PhD in data science, specializes in predictive algorithms and movie scripts.
“To find and bring Mark and Grace to Austin to join our team is really exciting for us,” said Landers.
Landers, who co-founded AUTHORS with David O’Brien in 2014, initially launched the startup to use AI technology to connect writers with publishers who would be most likely to pick up their books for publication.
When they hit the scene, Landers said the end goal was to take that AI platform and offer it as a separate algorithm for publishers.
By making the tech a free-standing platform, Landers said their brand of machine learning could also be applied to more robust opportunities — like scanning thousands of manuscripts for specific keywords that might predict market success rates.
However, Landers said they couldn’t quite get there until they proved the accuracy of their initial matching algorithm.
“When we were right over and over again, the power of the tech was realized,” said Landers.
AUTHORS will remain as is, supporting both sides of book deals between authors and publishers. Meanwhile, StoryFit’s machine learning technology will specifically identify what manuscripts have high potential for success.
“This is to offer a clear differentiation of our offerings — we’re not changing the company or closing down the AUTHORS side,” said Landers. “We want to continue to serve the writers through the platform as well as develop these keen, industry-specific analytics tools.”
StoryFit currently has 10 major publishers and distributors testing the metadata and keyword recommendations. The team is also developing other modes of comparison derived from the publishing and studio standards, including indie appeal and genre fit.
Image provided by AUTHORS.me
Want to get in touch? Let us know with a tip or on Twitter @BuiltInAustin
