Monday, March 1st, 2010

Beta: “Read Alike” recommendations

I’ve pushed through a beta version of a new recommendation engine.

The “Read Alikes” recommendations supplement our existing automatic and member recommendations. “Read Alikes” are based directly on the members who have your books—the people who “read alike” you, or whatever.

So far, opinion is divided. Some members love it, and are getting great recommendations. Others report a parade of things they already know about. Is it quite consciously, however, a beta feature. It may be improved, or it may go away. Most likely, it will go away and be replaced by a better overall algorithm, with better tools for managing your recommendations.

Labels: new features, recommendations

10 Comments:

  1. wisewoman says:

    "Is it quite consciously, however, a beta feature."

    The first two words in this sentence should be switched.

    Looking forward to trying this feature out. Thanks!

  2. Doug Cornelius says:

    It's a nice tool. But….

    I have my whole households' books in my collection. The end result is that it highly recommends books for my kids instead of me. Since my toddler's board books are all over my house, recommendations for board books are all over the "Read Alike" tool.

    But it's not you , it's me.

  3. dr says:

    Yeah, agreed about the kids books.

    You know what would be a great feature? Is recommendations *by collection*. I have all my kids books in a Children's Books collection. I would like to get recs for that, but separated from my own recs so I can browse each list at the appropriate time.

  4. Eliza K. says:

    Every "recommendations" feature I've seen has the same problem: it doesn't take into consideration ratings. Until there's an algorithm for that, quite frankly, I don't find much use for the feature. I find it more helpful to just look for my favourite books, see which members have also rated them 5 stars, and look through their collections. More complicated, but with better recommendation results because it's based on highly-rated by me books not on every single book I've ever owned or read.

  5. YavorD says:

    It is funny sometimes 🙂

    Model: The Ugly Business of Beautiful Women by Michael Gross NEW!
    96 copies. 1 reviews. Average rating 2.87. Recommended Mar 4, 2010.

    because of

    Mathematical Modelling Techniques by Rutherford Aris

  6. Martin says:

    It would be awesome to exclude works from certain publishers–my LT includes lots of superhero comics, and since it seems most people with comics only or mostly read comics, they flood my recommendations–without them, I could hopefully get some proper book recs.

  7. Emily says:

    I use LibraryThing for recommendations really often, thanks for continuing to develop ways that I can find new books!
    The terminology on this feature is a little confusing to me – I generally think of "readalike" as referring to a book – as in "Here are some books which are like book X."
    I agree with previous commenters that this feature would be more useful if it could be focused by the user – by collections, by tags, by publisher, etc.
    I also find recommendations more useful if there's a description of why the book is being recommended.

  8. Anonymous says:

    From the perspective of a member who catalogs only favorite books in my library in order to find something new to read, this feature is a nice addition. Sure the list does includes a lot of books I have already read but that fact only strengthens the likelihood that I will enjoy other books on the list. Thank you for continuing to develop new tools like this.

  9. Anonymous says:

    It would be great if members could choose which collections to include for the recommendations. Many of the books in my collection are for research and not necessarily for pleasure reading.

  10. Luchtpint says:

    Keep up the good work ! Some automated recommendations I receive don’t amount to much, really. Usually, I tend to edit this recommendations list afterward with the ‘No Thanks’ option to delete the useless ones. Author read-alikes may turn out to be a good solution for this in the long run.

Leave a Reply