From a literary agent’s webpage:
Manuscript Wish List:
- Adult and YA fantasy and science fiction except dystopian. In SFF, prefer upmarket writing and settings/characters that think outside SFF tropes with earthy female-based characters. If submitting urban fantasy, please no demons, vampires, angels, or werewolves.
- Literary fiction, especially magical realism that highlights particular culture or surreal intellectual humor, please read this post on what is magical realism before querying.
- Historical, if the voice is strong. Please no WWII narratives.
- Genre romance
- Cozy mystery
Almost all agents have something similar, i.e., a list of quite detailed preferences for the types of writing they will and will not represent. This one will rep historical novels as long as they’re not set in WWII. There’s probably another one out there who *only* reps WWII historical novels. Now, as readers, we all have our own preferences. I think postapocalyptic stories are pretty stupid and I avoid them. But if I were an agent I would totally still rep writers of postapocalyptic stories. The point is to make money, right? You don’t have to like the works of the authors you’re repping to make money off of them. Is this one seriously telling us that if Stephen King approached her with a WWII or dystopian novel she’d turn him down on the grounds that she doesn’t like that kind of novel?
One might say, yes, an agent can make money off any type of book provided only it sells, but given that you have to read all the books you rep life’s too short to read dozens or hundreds of manuscripts you’re pretty sure you’re going to hate. That raises the question, Is it necessary to read manuscripts in order to rep them? Does Stephen King’s agent read each of his new manuscripts and decide whether she’ll keep him as a client based on whether she enjoys them or not? And if not true for King, why should it be true for any other? The same method she used to judge King’s commerical viability she can use to judge anyone’s, just on a smaller scale
If an agent is predicting how many copies a book will sell based on her own response to the book, or even on her own judgment of market conditions, there’s a disconnect between the thing being predicted and the method generating the prediction. It’s like if one were to predict the presidential election by looking at the lengths of the cigarette butts found in the street*. There’s no real relationship between the data being collected and the thing being predicted. It’s just intuition, which is to say, decisions about one thing based on completely unrelated things. The literary agent trade is ripe for Moneyballing, for sabermetricking. One might not do any better, but I’m confident one could do just as well with one tenth the expenditure of time and effort.
(One tenth, that is, of the time and effort they say they exert. It’s possible they’re already using these methods and just not telling anybody)