reality check
Over on Scalzi's blog, in the comments, someone said:
"Assuming the publishing house knows something about what sorts of books sell in what numbers for what price, and it’s read the manuscript to be acquired, it can’t really be said to have taken a risk, can it? I would think the risk is especially low if you acquire and publish enough books for some kind of statistical distribution to kick in; so I aquire the rights to 10 books, 4 sell about as well as I expect, 2 sell a bit better, two sells a bit worse, 1 sells really well, one disappears."
I gave a shorter response over there (much of it duplicated here), but I'll unpack this a bit:
"Assuming the publishing house knows something about what sorts of books sell in what numbers for what price" -->Yeah, they know something, but it isn't much. The problem with books is that they're not fungible. They aren't bananas, or even fruit. You can't say, "Well, we've found that 20% of consumers like bananas, and 50% prefer oranges, so let's get more oranges to sell."
Because what happens is that next year someone introduces a new kind of banana, but you don't have a source for them, and you have to wait until they can supply some. And at the same time, everyone gets bored with oranges, but you already have a zillion oranges on the trucks from Florida headed to your warehouse. And then suddenly the fruit-buying public develops an insane mania for gooseberries, and you don't even know what the fuck a gooseberry is! how the hell are you going to find any, or know what markets want them, or how much to charge for them?
Also, suddenly no one wants sweet fruit. They want you to supply tomatoes, but you thought tomatoes were a vegetable? Why are they looking at you like that?
"I would think the risk is especially low if you acquire and publish enough books for some kind of statistical distribution to kick in"
"I would think" is always a dangerous phrase to start a sentence with. It is semantically equivalent to "I imagine, but I actually have no idea." The reality is that no, there is no statistical distribution that kicks in. The market changes without warning and much faster than you expect.
The prospect of always chasing the market is idiotic. A publisher makes money when they are the first company with the Big New Thing. Sure, the Harry Potter books inspired a renaissance of the YA fiction market (and specifically YA fantasy), but no one else is making JK Rowling levels of money, are they? Plenty of authors and publishers are doing very well in the market, but they are enjoying the momentum, not creating the force that moves it.
Westerns and horror are, at best, now niche markets (or have been subsumed into other genres). "Confessional" novels don't exist anymore. Publishers had to take losses on a number of such books because they were already in the pipeline when the market for them died.
As far as I can tell, huge swathes of the publishing industry really have no idea what will sell or what will tank, particularly nowadays. Even the editors with the best track records, with the full support of their marketing/sales departments, still sometimes have money-pissing duds. The equation of "what makes a book sell" is multifaceted, and full of variables and unknowns.
There are three ways in which your dependable operating income is made:
1. Backlist. Having books that reprint and continue to sell for years is like printing money. However, in the world of ebooks and piracy, this income stream is threatened. I'm unconvinced it's dead or even dying; more like going through a major life change. It will be a different situation when the transition to a stable ebook market is finally achieved, but in the meanwhile everyone is very confused and not sure what they should be doing.
2. Dependable bestsellers. Publishers make plenty of money with the books from reliably bestselling authors--I'm sure George RR Martin's publishers are confident the next SOIAF book will do well for them. (Note, though, that misreading this market can be extremely expensive for a publisher. Authors sometimes have downward arcs on their careers, and a publisher who enticed an author away from another publishing house by paying through the nose will often get burned.)
3. Unexpected bestsellers. The real money is in the book that a publisher paid very little for, but then takes off and they're scrambling to reprint it fast enough. This is a nice scenario that makes beaucoup bucks for everyone involved, and allows the author to leverage their next book as if they are in category 2, above.
Note that category 3, here, is the place where the publisher is not chasing a trend. This is what they all want to have, because the return on investment on an individual title is HUGE. But it is, by definition, the thing that can't be predicted.
The problem, as it has always been, is the midlist. These are the books that may or may not pay for their own overhead. They may or may not contribute a little bit to profit. But they are necessary to a real publishing house because (a) a publishing house only exists if they actually, you know, publish books and no one can own all the bestsellers; and (b) this is where the dependable bestsellers with backlists come from. In an ideal scenario, a midlist author gains a following and the publisher throws more marketing dollars at their books and helps push them up the bestseller list. Meanwhile, the author's backlist is still being published by the same company, so they're cleaning up in multiple revenue streams.
But the price of that is the high risk.
"so I aquire the rights to 10 books, 4 sell about as well as I expect, 2 sell a bit better, two sells a bit worse, 1 sells really well, one disappears."
Well, no, these numbers are totally mistaken. Out of 10 books on a typical mass market novel publisher's midlist, one will be a good money earner, two will be somewhat profitable, two will break even, and five will actively lose money. The problem for the publisher is that they don't know which five.
People seem to think that publishers have some sort of magic understanding of what will sell. If they did, we'd all be making a lot more money.