How Does She Do It?

I've mentioned that I have an assistant who helps me with my submissions. We're still early in this process, so she's a little unsure about which pieces to submit to what markets. She is still looking to me for guidance, asking me "Which one do you want to send to Kidney Quarterly? How about to Hubcap Review?" Normally, I think about what pieces she has, which of them hasn't gotten enough exposure lately, and make my answer based on that.

When you look at the submission guidelines for most magazines, if they're not genre, they're depressingly similar. "Furgpickl Monthly accepts only the most outstanding fiction. We're looking for evocative, edgy, brilliant writing that makes the hairs on our arms stand up." Or something to that effect. Something that says "Don't submit anything written on the back of a napkin unless you're already an established name, in which case we're already bribing your office assistant to mail us your garbage."

Every magazine suggests that you look through their back issues to see what kinds of stories they've published in the past, but I maintain that's bullshit. Here's the thing: even at the smallest magazine, you've got at least two - more likely three to five - people making the editorial decisions. If you can find even one story that's like yours, chances are that not everyone on the staff liked it but they put it in anyway. Now, you don't know what it is that they liked about it. Perhaps it had a dog in it with the same name as a particular editor's dog, and that made him feel kindly disposed toward it. Nevermind that the meat of the story was the one-legged woman determined to run a marathon. And your story about the one-armed tennis player has no dog and is therefore going to be seen as derivative and tacky. You are being asked to look into the minds of an unknown number of people and guess, based on choices they've already made, whether they would choose another thing that you've produced.

There's a term for this: it's called "collaborative filtering," and it was the common way that people did things back before everyone shopped online or in big, anonymous chain stores. You would go down to your local grocer and he'd say "Howdy, Myrtle. I see you're buying some peaches. You know, Frances was in yesterday buying peaches and she bought some maple syrup to put on them. Said she's been having them that way for years and she loves them." You'd buy yourself some maple syrup, remembering that time that you had a nice salad at Frances' house and she'd put raisins in the salad and that was really delicious too. The grocer knows that you know Frances and knows that you like her cooking, and is passing along her preferences to you in order to make an additional sale.

Places that do a lot of online business do this with complicated algorithms. Amazon.com does it - you see a window every time you choose a product that tells you that "People who bought a cheese straightener also bought a left nostril inhaler." Netflix does it, first asking you to rate a lot of movies and then recommending more movies to you based on the movies you rated highly.

Both the Amazon and Netflix systems are flawed, though. They can only be binary in their calculations (okay, so Netflix's system has four stars, it still comes down to "like/didn't like"). They can't drill down on why you liked or didn't like a particular thing. Sometimes, they make some spectacular mistakes and we look at something that Netflix or Amazon has recommended to us and think "What in the hell are they thinking? Why on earth would I want to buy a crossbow, just because I bought some cheese popcorn? Why would I want to see Jesus Christ, Vampire Slayer just because I liked Jesus Christ Superstar?"

But Amazon and Netflix are dealing with exactly the same basic problem. They're trying to discern, based on previous choices, other things that you would like. They don't know you personally and can't tell why you might have liked a thing in the past, so they're making their best guess. And remember, their "best guess" involves engineers having spent thousands of hours and millions of dollars constructing, refining and testing algorithms that are supposed to more closely mimic the millions of decisions that go on in the human brain when developing a preference.

I don't have thousands of hours or millions of dollars. I have just me and the knowledge of what any given magazine has published in the past. And sometimes, I don't even have that. Is it any wonder that I look at this whole process as a crapshoot? So, in the absence of the kind of engineering power that large corporations have applied to this problem, I turn to wisdom that has helped people in my situation for decades now:

If you throw enough shit at the wall, some of it's bound to stick.