The Pace of Myke Cole, the style of Maurice Broaddus, the grit of Abercrombie, the darkness of Brent Weeks? Having fun with Booxby: a flattering excursion with machine learning.

The novel of ours that has been most widely read is Wunderkind by a large margin. So it was with some curiosity and great interest that I put the novel through Booxby’s new machine learning tool to see what would happen. Booxby is trialing this as a tool for writers to help us figure out which readers our work will most resonate with. The idea being that highest on every reader’s wish list is that they find books that complement the writing style of their favorite author or authors. Whether or not this is true remains to be seen, but this is the fundamental principle behind Booxby as far as I can understand it. It should not be a surprise that a writer like me finds this idea difficult; we have not had such tools before. We have not needed them until very recently. The art of selling books has always been the role of the book seller. Writers write. Sellers sell. Somewhere along the line, particularly in the brave new world of Indie publishing, everything changed. Or perhaps this was always a falsehood and writers whose work sells were always on top of their genre, their readers, the zeitgeist, knowing what will appeal and which will not. Understanding the market. Booxby uses machine learning to find links between writers based on writing style and therefore (assuming their methodolgy is sound) helping readers and writers connect to each other based on style.

I first put Wunderkind through the machine. Wunderkind was written most recently and as with all writers, our craft improved, our art grew more focused and the end result is, I believe, a better product all around than the works that came before it. We wrote Splice first. It is nearest and dearest to my heart, but not as polished as Wunderkind. So how would the AI fare looking at these two books?

Much like Splice, when we wrote Wunderkind we were aiming for something fast-paced and fun. A story that was full of action but also explored the deeper relationships between the characters, as well as showing their inner story arcs of personal development. We wanted it to be fun and fast and quirky but also meaty enough to be realistic from a character perspective.

Splice started similarly. The two main characters (Elliot Goshawk and Arthur Fortune) would be very different from each other, both possessing that which the other needed but both not fully connected with themselves emotionally in a way that would allow them to see that.

Booxby looked at Wunderkind and then gave us some genre options to look at. These are, according to Booxby, the top books that most closely reflect the style of Wunderkind.

When we look at the Moods that Booxby identifies we get the following:

For Wunderkind, I am pretty happy with that. It is lively and compelling but also dark. Here are the books most similar according to Booxby:

So we get the complexity of Iain Banks with the darkness of Brent Weeks, the action of Goldsmith and the page-turning compulsion of Anita Shreve. Interesting. How does that compare against Splice?

Flattering for sure. And clearly I need to read some Kenneth Bulmer and maybe if you are reading this blog post, you do too! (Which I think is the entire point.)

What about mood?

All of this was exactly what we were going for with Splice. The humor of Hiaasen, the grit of Abercrombie and the darkness of Weeks. Did we achieve that? That is really up to you to judge. It is also interesting to see how the comp novels changed between Splice and Wunderkind given that there were several novels written between them.

For now, this has been a fun excursion down the Booxby rabbit hole. Not to mention a flattering one. We will definitely be looking to see how we can use these insights to make our books easier for readers to find and enjoy.