[dpapathanasiou]

Exactly.

The concept of "publisher as angel investor" only works if you assume the old publishing model of printing, storing, and distributing physical objects.

Electronic publishing is a completely different cost paradigm.

[jseliger]

Not as much as you think: writers still need editors. They still need someone to help them market their books, since every minute not spent writing is a minute not spent honing their craft. Merely saying "get a blog" isn't enough. They still need to be able to eat. And so on. A lot of these analyses have been busily pointing out physical costs account for a relatively small proportion of the overall cost of a book.

In addition, you, the reader, still needs someone to help you sort through the vast amount of stuff out there that you could read. Publishers, for all their other ills, help do this to some extent, despite all the false positives and false negatives that get through (or don't).

Electronic publishing might eventually become "a completely different cost paradigm," but it sure isn't there yet.

[mechanical_fish]

What you say is correct but doesn't even go far enough. What authors need is an insurance company.

What risk is being insured against? The risk of spending six months to a year per book to construct books, and then discovering that the books can't make enough money to feed you. The scariest cost of producing a book has nothing to do with paper, or even with marketing or editing. It is opportunity cost: When you're writing a book, you're not doing anything else, and you may be very well throwing your time away.

Software startups reduce this kind of risk by taking money from angels or VCs. (a.k.a. "publishers"). Or they try to gauge product/market fit as quickly as possible by doing things like shipping early, charging money, listening to customers, A/B testing, etc. However, this doesn't work well for books. Books have a smaller market than software, are harder to prototype, are much harder for customers to try out (it takes a lot of effort to thoroughly evaluate a novel), are much less valuable if they fail [1], and have a much more mature, undifferentiated, and well-saturated market: Your software has to compete against about fifty years' worth of earlier software products, most of which suck, and 95% of which won't even run on the machines your customers own. Your book has to compete against William Shakespeare. Literally.

How does the insurance work? A publisher reads your first book (or the first three chapters and an outline, which I'm told is a fairly standard way to submit a book proposal) and then offers immediate, free feedback: Either they accept it, which means pledging to stake a bunch of money and reputation to bring it to market, or they reject it. If your first book doesn't get accepted, it's a very valuable sign: You need to keep your day job, and you need to either become a better writer, try writing something different, try a different hobby, or learn to live without actually selling any of your work.

---

[1] Everyone on HN knows that having worked at a failed startup is no big deal. It teaches you a lot. You get a lot of resume-enhancing experience. And your failed software experiment may have valuable pieces that can be extracted. But bad novels are literally a dime a dozen, and an unpublished novelist with three failed novels is not worth any more on the job market than one with no failed novels.

[dpapathanasiou]

Those are valid points, but the publishing company's value add is not as significant as you might think.

My startup is working specifically in this area, and the results might surprise you.