Shepherd Launches New Book Recommendation Tool

Shepherd has introduced a new tool that allows users to discover books based on what readers who loved a specific book or author also enjoyed. The system relies on data from 5,000 readers who voted on their three favorite reads from 2023 to 2025, resulting in a dataset of approximately 15,000 books. Unlike genre-specific algorithms, this approach aims to provide cross-genre recommendations. Additionally, Shepherd is developing a full 'Book DNA' application that will eventually pull in Goodreads history to deliver deeply personalized suggestions. The first beta version of this comprehensive app is scheduled for release in late January in the USA, though it will start as a basic version. Users interested in testing the upcoming features can currently sign up via a provided form.

Shepherd has rolled out a new feature designed to help readers find their next favorite book. The tool allows users to enter a book or author they already love and see what other books were enjoyed by readers who shared that preference. This method differs from traditional algorithms by not limiting results to a single genre, often producing interesting and diverse results.

The foundation of this recommendation engine is a dataset built from annual surveys. Thousands of readers and authors were asked to share their three favorite reads of the year. Specifically, the current system is based on:

• 5,000 readers voting on their favorite books
• Books read between 2023 and 2025
• A total of approximately 15,000 book entries

The team chose to keep the dataset small for the initial phase to effectively test different approaches.

Beyond the current recommendation tool, Shepherd is actively building a comprehensive application referred to as the Book DNA app. The goal of this project is to provide deeply personalized book recommendations by analyzing a user's existing reading history. The application is designed to integrate directly with Goodreads, pulling in a user's history to identify patterns in reading preferences.

The development team is targeting a release for the first beta of this application in late January. However, the initial release is described as 'pretty basic to start.' The project aims to solve the significant challenge of matching readers with books based on people who like similar titles, moving beyond simple genre categorization.

Shepherd is currently inviting users to participate in the development of the new Book DNA app. A sign-up form is available for those interested in becoming beta testers. This initiative is part of a broader effort to address frustrations with existing recommendation systems. The creator, Ben, expressed a desire for better recommendations based on personal reading history.

Shepherd has previously showcased its development progress on Y Combinator's platform, referencing past 'Show HN' posts. These posts document the evolution of the platform as it works toward its goal of delivering high-quality, personalized book discovery.