Publisher’s Weekly recently started writing about the widening gap between front-of-list titles and the midlist (that is, bestsellers and everyone else; “backlist” refers to older books who publishers have the rights to which still sell reliably). Publishers are increasingly paying debut authors less and frontlist authors more.

Here’s the goddamn problem: nobody knows how to sell a book.

I’m not being flip. The average traditionally published book sells 3,000 copies. Self-published, less than that. Self published sans marketing push? You’re looking at single digit sales. Thing is, while marketing can help you sell a book, if publishers and marketing teams knew how to sell a book reliably every single time, they’d do that every single time. Publishing’s a small industry, fiction’s a tiny cross section of it, and books as products aren’t fungible– if you buy The Velveteen Rabbit and find its text replaced with Nabokov’s Lolita, you’ll be very upset even though you got “a book.” That makes them really unusual as products, and means what’ll sell one book won’t always sell another, even if it’s in the same genre.

I work with financial and business processes wonks, who are very bewildered why it takes so long for my agent to sell a book to publishers. I’ve told them it can take years, even if my book’s marketable, and they ask why. We work on a monthly and quarterly schedule in a very predictable industry.  An industry that leaves money on the table for any length of time is counter-intuitive.

I tell them that an imprint’s list– the books they publish for a season– is a cross between an art gallery (with curated content where the books are in conversation or part of a theme or what have you) and a portfolio of assets. Yes, you want reliable performers, but you must also take risks because authors, like assets, age. They go out of style or they straight up die, at which point you need to be developing other authors to help you make up that lost income.

Here is where my financial friends get mad at me: the numbers on the costs side are mostly real (paper costs a certain predictable amount, overhead costs a certain amount), but the income numbers are a fabrication. Nobody can look at a book and say how well it’ll sell. It’ll come down to whether a handful of people (probably white, probably from an affluent background) can convince their bosses (probably older, white, and affluent) “get” the book. And that’s about it.

Financial wonks either think I am making this up or ask me how this business stays afloat. I reply that it’s a tiny industry and it’s not staying afloat. Presses struggle, go under, or get acquired all the time. Especially literary presses rely upon the patronage of a few wealthy people. That’s alarming to me. Who gets to speak and make art shouldn’t be determined by a handful of donors who may just change their mind one day.

I make it sound like people don’t read and that’s why the industry struggles. People read plenty. Books aren’t dying. But capitalism requires not just profit, but growth in profit, and that demand is a pressure that creates mergers. It make presses more risk averse. It creates pressure to invest less in the midlist, pay authors and editorial staff less, and to transition more and more key roles to perma-lance. It makes agents have to do work traditionally reserved for editorial assistants (for a percentage of ever-shrinking author advances). And, of course, publishing is dealing with a retail nigh-monopoly in the form of Amazon and Ingram, a distribution nigh-monopoly. Both are negative influences on labor in publishing.

Because publishing has become (maybe necessarily) more risk averse, increasingly, authors themselves are becoming brands. Not just in a Twitter sense, not just in the sense you must make yourself available to every Karen and Todd who think that being a potential book customer entitles them to your time and attention, an actual brand. James Patterson releases a book a week by “co-writing” with authors who he commissions to write an entire book under his name. The infamous Trigger Warning, a book co-written by a very prolific writer named William W. Johnstone and J. A. Johnstone, was, in fact, only written by J. A. Johnstone because William W. Johnstone is dead. Has been for almost a decade. But he’d written a lot of books before that, his name still sells books, and his estate (and the publishers he works with) didn’t want to give up the income.

That is the direction things are headed in: authors as brands which publishing can’t afford to let die because it either won’t or can’t afford to invest in its future. The old excuse for publishing terrible books (like whatever Bill O’Riley farted out this month) is that the frontlist pays for the midlist. Well, now nobody’s paying for the midlist. The “necessary evil”‘s all that’s left.

As my agent tweeted this morning, “Publishing loves to call itself an “art” when it’s making bad business decisions and a “business” when it’s making bad art decisions.” Since Trump was elected, publishing has pumped out poorly fact checked palace intrigue book after palace intrigue book. The Booker got split (!) between Bernadine Evaristo and Margaret Atwood, who, let’s be frank, got the award because a cash-flush Hulu helped her campaign. Publishing as a whole is failing to take business or artistic risks, and is floundering.