So Who Should I Read Next?

If You Like X You’re Gonna Love Y!

People laugh about the weird offerings that Amazon or other online book venues will list as recommendations. These algorithms are becoming increasingly sophisticated, but in truth, is there any chance that any kind of program is going to predict what I will enjoy, when my tastes are admittedly all over the map? People look at me like I’ve sprouted an accordion out of my nose when I say, oh, that I love Fort Minor’s Remember the Name as well as Shostakovitch’s 11th.  Or in the book world, that I include both Middlemarch and Bored of the Rings among my long-time favorites. In both cases for wildly different reasons.

But we keep trying to find, and to recommend, things we like to people we think (or hope) might be receptive. I’ve seen, and loved, blogs that recommend older novels to readers who liked specific popular new ones, as well as the other way around. In the interests of that, I invite readers to offer some “If you like X, then you might try . . .” and give reasons.

In a discussion like this, if the kneejerk “Your taste is in your gymsocks if you like [whoever]” can be avoided, sometimes we can learn something about our own tastes, if not refine our search for new Good Stuff.

I’ll start by noting successful recommendations that I’ve made—no use in reporting non-successes as that can lead too quickly to acrimony, and also, I realize that though I might have had success linking some of these authors, another person probably hit a wall between readers they know. Which leads right back to crazy algorithms.

So. For purposes of discussion, and to keep myself from rambling on for hours, I’m setting the timer for three minutes, and I’m going to free associate as I recall successful recommendations that I’ve made, or were made to me in the form of “Oh if you liked that, then try this!”

Here goes:

I’ve had pretty good success recommending readers who wanted a reading experience similar to Katharine Kerr or Kate Elliott or Robin Hobb to Michelle West, Janny Wurts, Carol Berg, Joshual Palmatier, and Chaz Brenchley. A handful of readers loving the above came back to thank me for recommending Kari Sperring’s first novel. I’ve also had some success linking readers who loved the above and Octavia Butler to Carol Berg, and along another path, to Tiptree, and along a third, to Nalo Hopkinson.

Brandon Sanderson, David Coe, David Farland, have matched up pretty well, and sometimes those readers intersect with Robin Hobb, which has made me wonder if sometimes the readers don’t know there’s an authorial gender divide.

I had a tough time making recommendations to those who’d read and loved Greer Gilman until I suggested John Crowley, Theodora Goss’s short work, as well as that of L. Timmel Duchamp, and Susanna Clarke’s Jonathan Strange and Mr. Norell. Bingo!  Duchamp branched to Carol Emshwiller and Kelly Link, and a couple of times to new writer Sonya Taaffe, and thence to Shweta Narayan.

As for readers who wanted to venture into genre waters after reading Clarke’s novel, my success links were Neal Stevenson, Ian McDonald, and back to Greer Gilman. John Crowley linked successfully to Ian McDonald, and to Thomas Pynchon, but not from Pynchon to McDonald or Crowley. But yes Pynchon to William Gaddis. Weird, that.

Mercedes Lackey and Melanie Rawn have crossed successfully back and forth with Sharon Shinn fans, with a sidestep to Patricia McKillip. Also, which surprised me, the fans of the first two writers in this graph also liked Patricia Wrede, though I’d thought of her as leaning toward the more humorous writers, but that’s really only a couple of her series: Mairelon and the Dragon books. With fans of all of these I’ve had signal success recommending Roberta MacAvoy.

Re cross-genre, I’ve had good luck matching Patricial Wrede’s Mairelon duo and her three books with Caroline Stevermer to those who like Georgette Heyer and Regency romances, but the only success I’ve had with Jane Austen readers (few of the ones I know willing to try fantasy) has been Susanna Clarke’s book—two readers, and a third had already read and loved it. But once I scored by directing a hardline Austen reader to Ellen Kushner. From Kushner I’ve linked with happy results to Emma Bull, and in another direction, P.G. Wodehouse and Rudyard Kipling.

Also outside of the SFF genre, I’ve had good luck with Austen readers by recommending Patrick O’Brian, Dorothy Sayers, and Dorothy Dunnett. Re the latter, I’ve had some luck mentioning Guy Gavriel Kay and Diana Gabaldon, but of whom were heavily influenced by Dunnett, so maybe it’s not surprising there’d be a readership crossover.

Re a certain kind of voice in thumping action fantasy (Zelazny and Jack Vance being two of the older names in this discussion), I’ve had good luck with Patrick Rothfuss grouped with Steven Brust and Scott Lynch, and also Terry Pratchett, but Pratchett, like Lois McMaster Bujold, has reached across a wide range of readers, as has Megan Whalen Turner; the latter is a YA author, and I’ve been avoiding getting into YA here as that would be a whole nother discussion, but I’m keeping Turner there because these have been adult readers I recommended her to successfully.

What about Ursula K. LeGuin? I’d had the best luck recommending Eleanor Arnason, Judith Bear, and Rosemary Kirstein, again whose work is in dialogue with many of Le Guin’s ideas.

There goes the timer, so I will stop here, resisting the impulse to add another 473 names.

How about you? Who’ve you successfully linked for readers?

Sherwood Smith



So Who Should I Read Next? — 23 Comments

  1. Wow, you’re good at that.

    I’m pretty much only good at recommending for certain friends, whose tastes are familiar enough to me that I know (within a certain margin of error) what to point them at.

  2. I’m only good if it works! 🙂

    I think it’s more like I’ve been talking books for so many years that patterns emerge, and it’s a matter of remembering (ever-branching) patterns.

  3. This is fascinating, especially where one author can point to another, but the reverse is not true.

    Even without having read the majority of the writers (*shame*), I have heard enough about their works online to be able to see how and why your recommendations might work. What a cool thing–Venn diagrams of author likes. Especially, as you say, if a person picks up on one element in a work, and that’s what they like, and not some other element, which is what other people like.

  4. At least you’ve given me a list of names to investigate . . . so Thanks!

  5. Asakiyume: it’s interesting to discuss similar and dissimilar elements, especially when it comes to tastes. Like I will lose interest in something that shows no humor, but everything else is on the same level as a book I loved. Or, I’ll like one so much that I don’t mind the lack of humor (rare, but happens).

    Pilgrimsoul: Only fair–you’ve steered me to some good historical sources!

  6. It’s a tough game, but I wonder if taste really is as difficult to map as we think it is. Most book recommendation algorithms are fairly simple, and they’re based on one of two things: they either look at all the members in their database, find the members whose collections most closely match your own, and then find the books most of them own that you don’t; or, similarly, on sites like Amazon, they compare your purchase history and browsing with the purchase history of others and tell you that People Who Bought This Also Bought That, So Maybe You Should Check It Out.

    Netflix, on the other hand, has a much more robust algorithm that’s based not just on what you’ve seen, but how you’ve rated what you’ve seen. It looks not only at other members who have watched a similar set of movies, but at how those members have rated them, and whose ratings most closely match yours. At first, its algorithms are only moderately useful, but once you’ve rated a few hundred movies, it’s uncanny how accurately it can predict not only what movies you may like, but also how you will rate them. It’s occasionally way off, but far more frequently very close to right, and I can’t help but wonder whether if a book site got not only as much traffic as Netflix, but also had as many users actively rating all the titles they’d read, it’d be able to parse taste in books as effectively as Netflix has taste in movies.

    Which is to say: I think the problem with book recommendation algorithms is largely less that book taste is particularly difficult to map, than that books don’t enjoy as large and overlapping an audience that congregates as centrally at a few particular sites as films do. I think if Amazon, for instance, implemented a better algorithm and played up the use of rating the books you read, its recommendations would get a lot more interesting and correct very quickly.

  7. Interesting thoughts, Mastadge! (Though I have to admit that Netflix for me is more often wrong than right, so far. But I’ve only been using it a year.)

  8. I feel like I’m writing “I miss Alexlit’s Hypatia” all over the Internet lately, but it was a great program that I’d love to be able to use again, so: I miss Alexandria Digital Literature’s recommender program Hypatia. Its recommendations were always at least worth reading, and I found out about some amazing books through it. Since it was work-based rather than author-based, it was good for the cases of “if you like X, you’ll like this work by Y, but not necessarily anything else they’ve written”.

  9. Mary: I saw a similar sign in the late seventies, “If you like SWORD OF SHANNARA, try LORD OF THE RINGS.” In spite of friends passionate howls, I hoped it worked!

    Castiron: I liked Hypatia until, no matter how many ratings I did, it kept throwing two names at me that I really did not like, so I gave up. The idea was splendid, though!

  10. Lol, my tastes are so over the place that I’ve pretty much broken the Amazon system; and I don’t bother to go back and review / rate books I’ve bought.

    Not to mention that system always offers me German translations of books I bought in the original language. Hello, why should I be interested in translations when I don’t ever buy them (except for Russian novels)? 😉

    I use review blogs and fourms as guide what new stuff to try out. After a while one gets a pretty good grip on the personal tastes of the members and which recommendations may work.

    These discussions kept me off Rothfuss, for example, because while he gets lots of good reviews, it also became clear from those reviews that the books are a coming-of-age story with lots of small stuff going on. Not what I’m looking for in Fantasy (sounds like Harry Potter meets Der Grüne Heinrich, and I never finished HP and hated DGH which I *had* to read). Amazon keeps recommending it, of course (I think it’s because I bought Sanderson’s Way of Kings). On the other hand, I had been wary about Adrian Tchaikovsky’s books because of the insect kinden concept – didn’t want to read another Kafkaesque bug story – but learned in one of the forums that the kinden concept is more like a way they use magic and cool things like some kinden have a knack for technolgoy while others have fast reflexes instead, or a retractable sting, or can breathe under water (while the more technically adept kinden would build submarines), and that the books are a mix of steampunk and traditional Fantasy. Turned out they *are* a fun read indeed.

  11. The thing that has confused the Amazon recommendations for me is that I have bought books as presents for other people. I therefore get maths and programming recommendations that are way out of line.

    However, it was an Amazon recommendation that introduced me to Jasper Fforde on the grounds that people who liked Malcolm Pryce might also like Fforde’s books. (I’d stumbled on Pryce’s Aberystwyth Mon Amour in a bookshop in Aberystwyth and was intrigued enough to buy it.)

  12. Gabriele: oh yes, the language confusion.

    Helen: I hadn’t thought about that, but you’re right–when you buy books for others, oops, there goes the algorithm.

  13. FYI for folks who get a lot of wonky Amazon recs due to gift purchases and such — you can mark previous purchases as items not to be used for recommendations, and you can also rate specific recommended items as “not interested” if they keep popping up: (If you actually rate items on Amazon that’s also supposed to help further customize your recs along with your purchasing history.)

  14. Wow, you’re good. If I’m recommending books based on a “if you liked that then you’ll like this” scale (rather than a “here I like this book and happen to have it here and you’re bored” scale) I do it based more on series than on authors. I also tend to recommend books to people I know well, so I take into account their entire book history, and not just one author they really like. (Not to say I haven’t done the latter, just not as often.)

    I was talking about Robin McKinley last night with my roommate, so I’m curious, who would you match her up to?

  15. Presents can really throw off the Amazon recommendation system. For a while, I got recommendations for every motorcycle book under the sun, because I had once bought a motorcycle travel guide for my Dad. And ordering vaccuum cleaner bags for my neighbour (the neighbour has no internet and the vaccuum cleaner was so old that bags were no longer available at regular shops, but Amazon still had them) brought down an avalanche of recommendations for household wares, even though I have never bought housewares at Amazon before and never did again.

    But otherwise, Amazon’s recommendation system is fairly accurate, even though my tastes are ecclectic. Several times it has recommended books and DVDs that I was planning to buy anyway. I don’t rely on it thought, except occasionally as a way of discovering a new author writing in a genre I like. I get the recommendations for German editions, too, but then I have bought German editions of books I already own in English as presents, so that might mess the algorithm up. Oh yes, and it recommended that Sarrazin book to me, even though I have never bought any kind of political non-fiction there before and most certainly won’t buy that book.

    Otherwise, I’ve occasionally had luck with recommending books to other people, if I know them well enough. I can usually predict what books my Mom will like and she trusts my recommendations. Actually, she says, “Pick a book for me.” I’ve also supplied the neighbour kid with lots of YA fantasy, usually with good results. But my Dad has never read any of the novels I gave him, even though I took great care to take his interests into account. He used to read the Hornblower books, so he should like Patrick O’Brien. He used to build cruise ships, so the collection of funny short stories set aboard a cruise ship should work. And he’s been to Cuba a few times, so the crime novel set on Cuba should work. But he never read any of them, so now I only give him gardening and motorcycle books (and break Amazon’s algorithm by ordering them).

  16. Danceswithwaves: I’d need to know more about the recipient’s other tastes, in particular to find out which Robin McKinleys to recommend.

    But Blue sword I’ve recommended successfully to people who liked Tamora Pierce, Jane Yolen’s YA, Sharon Shinn, Georgette Heyer, Juliet Mariller, and Sally Watson, to name a few.

  17. But you can choose to ignore an individual book on Amazon for recommendations – not that that makes it perfect or anything, but at least it does stop it from recommending books for me based on my daughters’ university texts! Mind you, I’m a tiny bit protective of Amazon’s recs because that’s how I found Crown Duel. And the rest, as they say… 🙂

    I’m again embarrassed by how few of these authors I’ve read – or persevered with, in the odd case – but I think you’re spot on in the elements – er, element. And branching off from that is the fact that, as you say, an element that you *think* is crucial to you – or a dealbreaker – might on occasion be made up for by enough other elements being either present or absent.

    I don’t think I’ve ever really disliked a book you’ve recommended, but it happened with Connie Willis (she wrote the blurb for the book, rather than writing a rec on a blog), and it really puzzled me. And then an LJ/Goodreads friend just bounced hard off a book highly recommended by MWT which puzzled us both. I’m also kind of confused at fans of an author I love(d) raving about her latest books – it seems to me that loving her earlier stuff would almost *preclude* loving a few recent ones. It’s true for some of us but not for most, apparently.

    Looking forward to the YA discussion!

  18. Hallie: very true.

    Not all the authors I mentioned are favorites of mine, but I’ve read enough of them to gain an impression . . . in a couple cases, the very element that put me off might be an element I know someone else likes, and hence the recommendation.

  19. As for bouncing hard off a book recommended by a favourite author, this happened to me lately when a writer of humourous contemporary fantasy whose books I adore recommended a series as “the closest thing to my books there is”. So I bought the series in question and though the books started out as a funny fantasy, it took a turn towards darkness complete with explicitly described torture scenes halfway through. It was not a bad book by any standards, just not what I had expected at all.

  20. I’ve had very little luck recommending Dunnett to Gabaldon fans. Three times I’ve had the book returned after a chapter with the complaint that ‘it’s boring’. I suspect they felt the lack of a female first person character to identify with.

    But Good Omens by Terry Pratchett has always been a winner.

  21. Longtime reader (of your books :)), new reader to your blog!

    However, I am giddy that you’ve listed a pile of names for me to leap upon, just as I’ve finally made the trek back to my neighborhood library, one that I’ve neglected to visit for years. I’m rediscovering the joy of BORROWING books and I’m so happy there is more to look up now, other than to go off my Amazon wishlist of stuff that looked interesting. 🙂

    So thank you thank you thank you!!!