How BookTech winner Kadaxis is bringing better data to books

How BookTech winner Kadaxis is bringing better data to books

The latest in our series asking startups to share the challenges they face and lessons they learn as they grow catches up with Chris Sim from Kadaxis, winner of the BookTech Company of the Year Award at the FutureBook Conference 2016.

Winning the BookTech Company of the Year Award last year (following a live pitch-off and judging at the FutureBook conference) was a really valuable experience for the Kadaxis team. It gave us some great exposure with UK publishers; we made several good connections on the day and also with companies who reached out to us afterward - some of which have led to new business.

For those of you who don't know, Kadaxis is a New York-based startup which I founded in 2013. We use data science to improve book discovery across the publishing value chain, for authors (at AuthorCheckpoint with metadata and marketing tools), publishers (with metadata optimisation and data APIs) and readers (at experimental book discovery site BookDiscovery). 

What have we been up to in the six months after we won? Well, we’ve expanded our keyword analysis offering to incorporate the learnings and feedback we've gathered from adding keywords to thousands of books. From this data we've built predictive models to help determine which books are most likely to benefit from keywords, increasing the return on publishers’ investment. This new technology has been useful for evaluating backlist titles and complements our keyword curation product.

Example keywords generated for All the Light We Cannot See by Anthony Doerr

In the process, one of our biggest learnings has been that a one-size-fits-all approach often doesn’t apply in publishing. Each publisher’s requirements differ, from internal metadata curation workflow to catalog composition, and in many cases we’ve had to tailor our approach to facilitate the best outcome.

The second, probably obvious, learning has been that while tech can be cool and interesting, the most important factor for a publisher is how many more books we can help them sell. We now rarely discuss our technology in much depth with publishers, and instead focus on results and practices that make a difference when using keywords. Publishers sometimes define their keyword methodologies based on hearsay or web SEO principles, so it’s been rewarding to share with them hard data on what works.

In terms of our own journey, it's been pretty much business as usual, without any major pivots. At the time of Futurebook, we'd already proven our business model and were generating revenue, so since then we’ve focused on expanding our client base, and improving the effectiveness of the product. We measure product success through several metrics which are all aimed at increasing search visibility (and then revenue) for our clients.  While it’s been tempting to explore new technology and product directions, we treat anything that doesn’t help us deliver on the original vision, as a distraction. Since Futurebook, we’ve doubled down on our efforts to improve the ROI of using our product. We probably won’t pivot too far from this capability in the short term, but might investigate other verticals in future.

The Kadaxis marketability assessment API generates writing metric scores

Our biggest challenge today is scaling personnel. There’s been increased demand for our services and keywords in general. While scaling our tech hasn’t been an issue, meeting the demands of day-to-day business operations and client management has been increasingly challenging as we’ve grown.

Another potential future challenge is competitors entering the space. We’ve developed the market and increased awareness of the benefit of keywords, and this hasn’t gone unnoticed by some fledgling publishing tech companies. In the past, other companies have dabbled with using machine learning to generate keywords, but to date there aren't any we're aware of that have persisted. Beyond solving the the technical solution, there's a huge amount of work required to understand the nuances of publishing book data, search engines and how individual publishers operate. It's challenges such as these that we'll be more focused on as the business moves into the next stage.