[meuk]

> very little of the cost of making a book is tied up in the physical format

Either that's bullshit, or the costs are entirely artificial. I really, really want "Concrete mathematics" by Donald Knuth. I really, really don't want to pay €80 for it. A good alternative for me would be the e-book, but it isn't any cheaper. I doubt if anyone will buy the ebook for that price while there are free copies available online.

[dgacmu]

It's entirely dependant upon volume. According to [0], the average US non-fiction book will sell under 3,000 copies in its lifetime. For the proceeds from that to pay the author, editor, and marketing... Well, that's not a lot of copies to amortize those salaries across.

The numbers are obviously very different for a New York times best seller, where physical goods costs will represent a larger fraction of the cost. But most books aren't NYT bestsellers, (even? Particularly?) When they're by Knuth on advanced topics. :)

you could call paying all of the people involved in the process of creating a book artificial, but if we don't, we don't get more books.

[0] https://m.huffpost.com/us/entry/us_1394159

[meuk]

You make a good point: There surely is a difference between someone like Donald Knuth writing a book that is relatively niche, and a best-seller. For one, I can imagine that a page of mathematics requires more expertise and time to write than a page of fiction.

However, there are dozens of free e-books on the web, written by respectable and knowledgeable professors, and this tells me that the only real cost involved is the effort of writing itself. The distribution and marketing cost for an e-book like 'Concrete mathematics' should be nearly zero. Unless my €80 goes directly to the authors, I'm wasting money. This is defendable for a physical book, but not for an e-book.

[ineptwriter]

Reacting to this: "I can imagine that a page of mathematics requires more expertise and time to write than a page of fiction."

I cannot say exactly how much time it takes to write a page of math, but I do know that many underestimate the time it takes to write a page of fiction. Sure Dan Brown probably knocks out a page of pulpy prose in an hour. But that's not exactly equivalent to critical reviewed, prize winning, literature. Nor is it what most writers strive for. Many fiction writers have spent years writing their books, recieved graduate degrees (some that they paid for), stuck failed novels in drawers, produced multiple drafts, been rejected hundreds of times, etc. I.E. labored over each page and fought for each step. Beloved was Toni Morrison's 5th book, not published until she was 56. That's decades of work to get to that point. Try looking at any page of James Joyce Ulyssess. That book alone spawned the PhDs of hundreds (thousands?) of scholars in English departments across the world just to understand it. Math too takes this level of work, it's just that they're more equivalent than some think.

Except, even though many fiction writers labor over every word and punctuation mark, there are not the same institutions to support them as there are mathematicians. Want to pursue a career in math? - enroll in a fully funded PhD in Mathematics, hopefully get a post-doc, then a tenure track job, etc. Admittedly not all succeed on this route, and it's not easy, there are teaching and publication responsibilities, but it exists. For writing? You're on your own until you produce critically acclaimed works, generally two successful books, usually while pursuing separate careers. Then you can apply for tenure-track jobs. There's not the same caccoon to usher in writing talent. Even Dan Brown, whose prose I just bashed, spent 14 years as a songwriter and High School English teacher before the Da Vinci code (his fourth book) garnered attention.

So I'd disagree, many (most? - not sure) fiction takes significant expertise and time to produce, and usually without the support of academic institutions that budding math talent has.

[logfromblammo]

To write a page of fiction, write ten pages of fiction, and then remove the nine that no one wants to read. Do all this entirely on your own, while working your day job.

To write a page of math, write a paragraph of math, illustrate it or insert the proof, typeset in LATEX, and leave the rest as an exercise for the readers. The readers will probably be your students, that bought your book because you required it for the course, and will make revision suggestions during class or office hours. Get a colleague to proofread for you, as you will proofread their math. Collect your pay for teaching and/or researching math while writing the math.

[meuk]

I have no doubt writing fiction isn't easy. I'm just saying that in general, a math book like I mentioned takes more time and expertise to write than fiction. Not in the first place because it requires that you have at least the level of expertise of a graduate student (which is 5 years of study). I'm not sure what Ulyssess being intensively studied has to with it. The only thing which makes a math book easier to write is that it's more accepted to distribute the work between different writers.

Some caveats apply: Many professors have teached a course repeatedly, at which they have such a level of mastery of the subject that they can just bang out the theorems and proofs (I've had one professor who wrote the 300 pages of lecture notes during the course). When the book actually contains new information, it takes much longer.

[ineptwriter]

My point is less a knock on mathematicians and more that folks underestimate the expertise and time spent by fiction writers. And there is less support for writers along the way. And the emphasis on "more" here illustrations that. People tend to discount fiction writers because they write outside of institutions and because popular books distract and detract.

Expertise: Take the most recent 4 Fields medalists. All 4 have PhDs (all of which were completed in less then 5 years, though that's more a testament to them) and worked in academia/research since undergrad. Compare them to the two MacArthur Genius grant winners who are writers (announced today), only one has a PhD, the other an MFA. So does that mean that the Fields winners have more training? Maybe, but there are very few creative writing PhDs in the world (maybe 100 PhD spots across all programs vs. what 5,000 math PhD students?). MFAs are considered terminal degrees, last 2-4 years, and many require you to pay for them vs. Math PhDs mostly are fully funded. So what's a fiction writer to do? Get a PhD in something else (Valeria) or string together fellowships and jobs for (4-5) years (in Ocean's case) before you can get an academic appointment. Does that mean Ocean's 2 year MFA + 4-5 years on his own is less expertise than a math PhD? Maybe, but there isn't the same support. After this, mathematicians and writers start looking the same (have academic jobs, teach part-time, work part-time). Except both writers produced at least two books on their own before they could get academic appointments whereas none of the Fields folks have left academia since undergrad. That's years the writers had to support themselves while also writing.

Time: Hard to tell on this one. But Alessio Figalli (Fields) has produced 125 papers that seem to be about 20 pages in length in 12 years of publishing, so that's ~2500 pages, or ~200 pages a year. And Valeria Luiselli (MacArthur) has published ~1000 pages in books in 8 years, or ~125 pages a year. If they both spent the same amount of time working and each page was unique (may not be the case for Alessio), then Valeria spent more time per page. If Valeria had less time because they had a full time job versus part time teaching or Alessio's papers repeat some of his work, them maybe it's similar amounts of time per page. And if a student leaves their MFA with half a book and a few small publications, keeps writing the book for a few years before publishing it (Ocean Veong's Night Sky With Exit Wounds) vs. a math PhD publishes their thesis plus 2 short papers out of their PhD (Akshay Venkatesh's thesis + 2 short papers from 2002) isn't that roughly the same output in the same amount of time?

Admittedly one can't use isolated examples to deduce general truths, but it's illustrative. And when you start comparing, say Donald Knuth to Toni Morrison, both have decades of expertise and spent hours laboring over each page, so it starts to seem moot. And fiction books tend to sell more copies so can be priced less because of scale.

[JoeAltmaier]

Coincidentally, it costs between $2000 and $4000 to print and market a book. So, what, say $1 per book pays for that. The rest pays the author.

But wait! Authors only get a buck or two royalty per book. So why do they cost so much?

Use your imagination.

[bregma]

> it costs between $2000 and $4000 to print and market a book

It costs between $2000 and $4000 to print a book. It costs about the same to warehouse and distribute a typical hardcover.

It costs maybe $100,000 to market a book. Often much more.

Printing, distribution, and royalties are a small part of costs in the publishing industry. Marketing and amortized overhead are a large part of the costs.

[greedo]

Most books, especially non-fiction have incredibly small marketing budgets. Often it's up to the author to self-promote the book.

[vorpalhex]

If you spend $100,000 to market a book that only sells 3k copies.. maybe you're bad at marketing.

[SolaceQuantum]

Marketing, type setting, cover design, editing, distribution channels, licenses for fonts, lawyers for contracts between everyone, accountants for managing the money flow of all the above, analysts to track how well each genre is selling and why, etc? And if the cover is particularly glossy, has metallic shiny lettering, velvet, etc. That also jacks price because that stuff all has to be custom made.

[TheCoelacanth]

The author's cut, the publisher's cut and the cost of printing the book are each about 10% of the gross[1]. The rest goes to the retailer.

[1] https://www.antipope.org/charlie/blog-static/2013/03/why-i-d...

[falcolas]

Many authors also get paid ahead of time to write their book as well. So, a portion of that also goes towards managing the loan given to the author, and distributing the risk of any given book not living up to its sales expectations.

[bovermyer]

Some of this is economies of scale. It's fairly cheap per unit to print 1,000,000 copies of a book. It's much more expensive (again, per unit) to print 1,000 copies.

Books like Lord of the Rings which are guaranteed to sell a large number of copies are much less of a risk to print large numbers of than books like Concrete Mathematics. The publisher rolls that expectation into the retail price.

Additionally, all one-off production costs (like paying an artist for cover art) are rolled into that retail price also, with the expectation that X number of copies will sell over the "initial life" of the book.

That "initial life" is usually the first year of a book's publication.

[AJ007]

The Doom and Wolfenstein 3D Game Engine Black Book(s) were priced at cost, per the author. Print on demand full color is very expensive.

[tbyehl]

> Either that's bullshit

It's misleading in that it ignores all the costs between creating a physical copy of a book and a buyer placing it in their hands. There's the obvious things like logistics and less obvious things like paying for placement and future liabilities for "unsold returned inventory" (typically destroyed instead of returned because an unsold book is less-than-worthless).

But 10 years ago, an e-book was often cheaper than a printed copy, allegedly because Amazon was willing to take a loss and everyone else had to follow to compete. The Great Steve Jobs didn't want iBooks to compete with Amazon by losing money, negative gross margins aren't the modern Apple way, so he colluded with the Big 5 publishers to force the agency pricing model on everyone. DoJ slapped them all on the wrists and the agency model lives on.

[diffeomorphism]

Concrete Mathematics is 672 pages, hardcover for $72 on amazon.

Lord of the Rings is 3 hardcover books, 1214 pages total for $39 dollars.

Considering that you get much more "book" out of the second one, the content seems to be most of the price.

Similarly, if you look at China-only copies of math books, you pay about $10 (IIRC), which gets you admittedly bad quality paper and binding but still makes the publishers money.