Why the same books follow you everywhere, and how to find hidden gems

Popularity is the most visible fact about a book and one of the least useful for deciding whether you'll like it. A book being widely read tells you it appealed to a lot of people. It doesn't tell you whether it will appeal to you. Yet almost every place you discover books ranks them by popularity first.

Why popular books are everywhere

Recommendation systems have a structural bias towards books that are already popular. The mechanism is simple. Popular books are in more readers' histories, so they come out as “similar” to more things, so they get recommended more, so they get read more. Each step is reasonable on its own, and together they form a loop. Researchers call it popularity bias. In systems that learn from what people click, it's also called a feedback loop.

The result is familiar: the same few titles on every list, table and “readers also enjoyed” row. That isn't because nothing else is good. Nothing else ever got the exposure to prove it.

The long tail

Plot books by how many people have read them and you get a steep curve. A few books have enormous readerships, and a very long tail of books have small ones. The tail is where most books live, and it's where readers with specific taste are most likely to find something that fits them better than any bestseller.

Match Book grades every book it reads on a four-step popularity scale. Across its catalogue of 1,407 books, about 13% come out as mega-bestsellers, 40% as well known, 35% as moderately known, and 12% as hidden gems. That catalogue is built from what readers import, which already skews towards the popular end, so the real tail is much longer than these numbers suggest.

Why average ratings don't help much either

Star averages look like a quality signal, but they're a weak one for choosing. On most book sites the averages of established books bunch into a narrow band. The difference between a 3.9 and a 4.1 is mostly noise, and it depends heavily on who chose to read and rate the book in the first place. A book beloved by a small, devoted audience can end up with the same average as a book that everyone read and felt mildly positive about.

An average also hides disagreement. A book that half its readers adore and half abandon can average the same as one everybody finds pleasant. For a reader deciding whether to start it, those are completely different books.

How to find hidden gems that fit you

1. Start from what you loved, not from what's popular. Describe your favourites by mood, pace, weight and theme. The taste guide shows how.
2. Search by those qualities, not by rank.A list of “dark, slow-burn books about grief” includes obscure books alongside famous ones. A bestseller list never will.
3. Follow similarity chains. Start from a well-known book you loved and look at what reads like it, then at what reads like those. Each hop tends to move further from the head of the curve. Every book page here lists its six closest neighbours for exactly this reason.
4. Try a debut. First novels are the part of the tail with the most variety. Here are the debuts Match Book has read.

Or go straight to the hidden gems collection. Every book in it shows its full reading, so you can check the fit before you start.

The Match Book app lets you turn popularity into a dial: keep your taste on every other dimension, and ask only for books relatively few people have read.