By Adam Fortuna
We’ve launched the first version of something we’ve dreamed of since our very first month: Match Percentage!
For every book you haven’t read, we generate a score from 1% to 100% for how much we think you’ll enjoy that book.
To start, this score is entirely using collaborative filtering. Collaborative filtering is a technique where we use every rating you’ve ever made to find other readers with similar tastes in books. We look at what those users have rated books you haven’t read, and voilà, Match Percentage!
Today this is a context-unaware filter. It doesn’t take into account the author, genres you love, or any other information. As we grow and learn more, we’ll improve this algorithm to use more data.
Here’s what this looks like in practice. It’s attempting to find people with similar tastes and predicting based on that.
Eventually, we want to show it wherever we show a book (or at least make it available). For starters today it’s available in a few places:
You’ll only see Match Percentage if you’re logged in. You’ll also need to have rated some books – the more the better!
For a book to have a match percentage:
In other words: the more books you rate, the higher the chance of seeing a match percentage for a book.
Right now this algorithm is based on a relatively small sample size. As more people join Hardcover, the algorithm will get better and better.
Match Percentage is pre-computed. That means that behind the scenes we’re calculating your possible rating for every book we can. It’s a lot of pre-calculation for many books you’ll never even look at.
We update match percentage every hour for anyone who was active during that time. If you interact with a book in any way, your scores will be updated within an hour.
We recommend you import your library from Goodreads or Storygraph then wait an hour.
If you just want to see what your match percentage is for a book, you can search for it and view the book page. If we were able to calculate a score, we’ll show it there.
When we talked to readers and users of other book platforms, one thing stood out: the process for adding books was very similar.
Readers would hear about a book, search for it, visit the book page and then save it to their “want to read” list for later.
But why go all the way to the single book page if you just want to save it from the search results? Hardcover makes it a step easier by showing the Match Percentage and an easy way to save it right from the search results.
The best way to use it right now is to head over to your “Want To Read” list and sort it by “Match %”.
This gives you an instant look at what books Hardcover thinks you’ll enjoy the most. These numbers will change over time as you rate more books and other people join Hardcover (and as we improve the formula).
Whenever you see another reader on Hardcover, you can head over to their “Read” list to see every book they’ve ever read. For example, I can head over to my wife Marilyn’s Read books. On there I can sort by Match Percentage and see which books on there I might enjoy the most.
I love having someone to chat about a book with once I’ve finished it. Choosing a book from someone’s list that they’ve already read is a great way to deepen friendships!
Want to find someone to try this out on? Check out the Supporters Page.
Lists are one of the core features of Hardcover. You can create lists with any number of books on them for any reason. This includes creating lists that exist outside of Hardcover.
For example, on our Lists page, we’ve created dozens of lists from popular places: NPR, The New York Times, Time Magazine, Condé Nast, and Pulitzer Prize-winning Novels.
For example, I love Science Fiction and Memoirs. I can head over to the The New York Times’s 50 Best Memoirs of the Past 50 Years or the Npr Top 100 Science Fiction and Fantasy Books of All Time and browse for books to read.
But narrowing down 50 books to the one I should read next is more difficult. That’s where Match Percentage comes in. If you’re logged in (and have rated enough books), you can sort these lists by Match Percentage.
Head over to our List of Lists to find some books that match what you’re looking for.
If you’re like many readers, your Want to Read List grows at a faster pace than your Read List. This list is there to save anything you might want to read someday.
I don’t revisit this page nearly enough. On Goodreads, it’s very difficult to explore it in ways that are useful for discovery.
One great way to clean up this list is to sort it in reverse by Match Percentage. This will show you the books we think you’ll like least.
When I did this, I was surprised to see Lord of Light, a Roger Zelazny science fiction book that Reddit absolutely loves. With a score for me of 54%, it’s one of the lowest. Now I want to read this one to test the algorithm and see if it’s on the right track.
On your Feed Page, you’ll see all kinds of activity from other readers. That includes answers to prompts, books added to lists, and (of course) what books people are reading, saving, and finishing.
You’ll also see a match percentage from any books we can calculate one for. I’m a lot more likely to investigate a book on my feed if I see that it has a very high match percentage.
Hardcover is bootstrapped, 100% free, and with no ads. We’re able to do that because we’re small. This feature alone will not scale very well.
In order to keep the lights on, we spent $9,207 in our first year to keep the lights on.
As we grow and more readers join our expenses will no doubt go up. We’re relying on our Patreon Supporters to make this project a success. If you’ve been waiting for a Goodreads alternative, I hope you can support and help build the next generation book social network.
While we’re only showing it in a few places right now, we can show it anywhere (and everywhere!).
We’re not even showing you the books that have the top match percentage across the site! That’ll come next.
If there’s a place you’d like to see it, or you have an idea on how we could improve Match Percentage, join our Discord and let us know.