Shepherd

What it does

Shepherd collects annual reading data from thousands of readers and authors who share their three favorite books of the year. Users enter a book or author they like, and the service shows them other books that were loved by readers who also enjoyed that same book or author. Results span across genres rather than restricting recommendations to a single category.

Who it is for

Readers seeking personalized book recommendations based on their existing reading preferences. The service appeals to people frustrated with generic recommendation systems and wanting discovery based on what readers with similar taste actually choose.

Pricing

The site does not show prices.

How it stands out

Shepherd's approach differs from traditional algorithmic recommendations by grounding suggestions in actual reader behavior. Rather than analyzing book metadata or text similarity, it uses crowdsourced annual lists from thousands of real readers. The cross-genre approach means recommendations can surface unexpected connections between seemingly unrelated books, rather than staying within narrow category boundaries. The service relies on direct author and reader participation in its annual survey process.

What a founder should check

Anyone considering a similar business should investigate several factors. First, examine how the existing book recommendation ecosystem—including Amazon, Goodreads, and major publishers—currently handles similar use cases. Goodreads in particular already has massive user bases, reading history data, and established recommendation features. Second, evaluate the switching costs and data portability issues. Readers often have years of reading history embedded in Goodreads; pulling that data to use elsewhere involves friction. A founder should verify what technical or contractual barriers exist to exporting reading histories and whether users see enough value in a new system to manually rebuild their profiles. Third, assess the long-term moat. Shepherd's advantage appears to depend on maintaining an active annual survey of readers and authors. Competing services could potentially achieve similar results through other data sources—mining public reading lists, using existing Goodreads data with permission, or applying machine learning to existing book databases. A founder should validate whether annual surveys create sufficient differentiation or if the core insight can be replicated through alternative data collection methods.