Editorial by Marc Santandreu, CEO, Tekstum
Tekstum is a startup based in Barcelona that offers insights to publishing firms that can help generate a greater demand for their texts. We analyze book reviews on literary platforms (like Goodreads or Amazon) and readers’ comments on social networks (such as Twitter and Facebook). We also measure the social engagement around a book and the author’s reputation.
From our point of view the publishing industry is characterized by several things:
- Intuition plays a large role on the decisions made in the publishing sector.
- The production model is based on supply and not on demand. This model is far different than ones used in other industries.
- The publishing industry doesn’t currently take advantage of all the data available on the internet.
Data available for books increases every day, but this information is not being properly analyzed and used. It is a significant business opportunity. Big Data is adding value and benefits to more and more sectors: banking, insurance, medical. Why not in the publishing industry? Is there no data to analyze? To the contrary, the Internet provides a never-ending source of data.
Publishing firms almost only work with quantitative data. By that, we mean sales. Sales numbers tell us which books are liked and which books are disliked. But they don’t tell us anything else. They don’t explain why a book is liked or is disliked. Publishers don’t have this information. By analyzing qualitative information, we can answer this question.
For this reason we built a complex algorithm based on Big Data and Artificial Intelligence in the sector of Natural Language Processing to discover the emotions and feelings that a book transmits to its readers, to transform it into useful data for the publishing industry. We provide real-time data for making decisions: more efficient and effective decisions.
This is because a book is exactly that: feelings, emotions, subjectivity.
Our tool allows publishers to know why a book is liked or disliked and also to know the reader experience, so you can focus the production model on demand. And then the qualitative information defines reader preference, which allows publishers to anticipate future trends and adapt to demand.
There are other sentiment analysis tools in the market. But most are not for books. Analyzing book reviews is not the same as analyzing reviews of hotels and restaurants. So with the help of Artificial Intelligence, Tekstum has built a sentiment analysis tool specifically for books. If you work in a short domain the tool precision is greater so the information that you provide to the publishers is much better.