I wrote Programming Bitcoin for O'Reilly and it was published 3 years ago. Some observations:
* You won't make that much money directly, but I have probably made a good amount indirectly. The credibility from having published an O'Reilly book is not to be underestimated.
* I got the animal I wanted (honey badger) by asking for it. I did justify the ask by showing my editor the Bitcoin honeybadger memes.
* The book writing process was a grind. Because my book was so focused on programming exercises, I wrote my own suite of testing tools to make sure everything was consistent. It felt more like a software engineering project than a bunch of blog posts.
* I have since introduced more people to my intake editor and they've published their own O'Reilly books.
* I don't get much of a discount on my own book. I just buy them through Amazon.
* They give you a framed picture of your cover when your book publishes.
That is insanely low. Like, "why would anyone even remotely consider publishing with O'Reilly" low. Even if they have a large audience reach, 10% should be rejected out of principle.
The economics of writing a technical book for a publisher are very poor, even with better royalty rates. There's an adverse selection problem: many of the people most qualified to write a book bill out at rates that make sinking time into a book project irrational.
It's a reason I take self-published technical books more seriously than "major" publisher books; the incentives are better lined up. 20 years ago, you could make an argument that working with a publisher would get your book on shelves at brick and mortar stores. But it's 2022 and to a first approximation 0% of technical book readers shop at Barnes and Noble.
The other reason you can imagine someone writing a "major" publisher book is that they lead a project they want to generate attention for. An O'Reilly book about your web framework has the benefit of implicitly endorsing the framework as something mainstream. That's a valid reason to do it --- for the author. But the incentives aren't lined up well for the reader!
Editorial effort and quality control also plays into the purchase decision: I have a sense of what "an O'Reilly book" will be today, just like I have a sense of what "a Packt book" will be. With a self-published book, I might know what the author is like but I can't be so sure what the writing and structure are like.
O'Reilly books today may not be the automatic promise of high-quality texts they used to be when I was writing (last millennium!), but I know that there's a reasonable amount of effort that goes into selecting titles and improving the quality of the book.
You can always find a self-published book that's awesome and an O'Reilly book that's crap, but my experience is that the median O'R book is higher quality than the median self-published tech book.
(I'm not saying you should only buy O'R books -- my library has plenty of non-O'R books)
When buying technical books I don't care particularly about writing skills and structure. I only care about the author credentials and skills.
I'll overlook a bad structure and even bad writing if the content is good and will give me value.
I'm way more likely to buy the book at 25% of the price with no editor than at 100% of the price with an editor.
A well written book is so much easier to read that I'll pay the premium (within reason). When I read a textbook by Hadley Wickham, I can tell how much he loves novels; he knows how to write in a way that's engaging for the reader. I've tried to slog through technical books that don't have the same writing level (despite the same technical proficiency of the author) and it's painful.
Enter "Independently Published" for publisher, then whatever keyword you want. Using "network" turns up a really good one I'm currently working through:
"Beej's Guide to Network Programming: Using Internet Sockets"
That's about normal for the publishing industry. 10-15% for hardcovers and 7-10% for paperbacks. It's also normal to be paid an advance against the royalties, so you may not even see any royalty checks if the book doesn't sell enough.
Sure having a book published by O'Reilly is a good brag point on a resume. I wonder what other IT book publisher could offer similar (obviously smaller, but still) level of recognition at higher royalty rate.
No Starch gives options for authors as far as advance/residual. Still not lucrative, but I think they have a strong brand and it's probably a good career move.
Lost Art Press specializes in fairly niche woodworking content. All authors get the same 50/50 profit split deal. Their books are printed with great care and quality, and the content is some of the best woodworking content available.
Of course that’s 50/50 of the profits vs 10% of the cover price.
Bookstores pay 60% of the cover price to the distributor who typically pays around 50% to the publisher.
Then there’s the cost of actually producing the book. And presumably things like editing, etc. are considered costs and not part of the publisher’s split. In the end, that 50/50 split probably comes out pretty close to 10% for the author.
Prag offers 50% of profit versus 10-20% of gross revenue shared by traditional publishing houses. It's often a little more money, but not as big a difference as they want you to believe.
Leanpub is a much better option these days in my opinion.
The trouble with profit-sharing is that however tightly written the agreements, you're at the mercy of the entity controlling the value chain, since they can cookie-jar any surplus out of existence.
The film industry is notorious for this, so much so that such sharp practices are named for them¹.
To be clear, I'm not accusing PragProg of anything of the sort, but on general principle I'd be super wary of any such deal.
Motivation for writing is not money. Making a living (in the US) writing technical books is highly unlikely. 10% is a nice surprise "mailbox money", not something you depend on.
Whatever the motivation, you want to best product you can deliver, right? That's where a professional publisher comes in. O'Reilly is best of the three publishers I've worked with. My editors were great, they tighten the prose so much and I learned a lot about writing along the way.
Oh, one more thing - translations. I get extra money when the book rights are sold for translation. And seeing your book in a different language is a unique feeling.
If finance is your only motivation for how you spend your time, do not write a book. It's a very long tail business. Good sales, whether you self-publish or have a publisher, come from great marketing -- the right topic and a high-profile author who is very visible and gets the book noticed by many people. The odds of that being you are very small indeed.
Why would anyone write a book? To spread the knowledge, to gain some mana for the author, to gain some mana for the product, to support a consulting business. ("Author of ...")
All those motivations are more strongly rewarded when you're published by "A Real Publisher" (doesn't have to be O'Reilly). Real publishers have distribution, and have a brand you and the product and the consulting business can benefit from.
If you think you have the reputation and the visibility that will sell your book, definitely think about self-publishing. If you're prepared to invest the time in things that aren't coding and aren't writing (managing reviews, hiring a copyeditor, getting figures, having a professional index built, designing or managing a designer of a cover, etc.) then self-publishing might let you have a commercial success of a book.
But books are like startups: most die in the dark. Self-publishing or going with a name publisher is just arguing about 10% vs 90% of bugger all.
Yes, you do pretty much need to pay at least something for editing and copyediting. (Also possibly at least cover design.) But you probably shouldn't be expecting a publisher to be doing heavy editing for you either.
As for the others, you also shouldn't be expecting the publisher to mail out review copies, arrange book signings, etc. Or other promotion that goes beyond just being part of the publisher's stable of books. That's going to be on you.
As for the others, with print on demand someone can order a book through Amazon. No involvement on your part.
The printing cost for a 200 or so page 6"x9" softcover in Kindle Direct Publishing with color cover and B&W interior is about $4. So if you sell it for $10, you collect about 50%.
However, in my case, I give away an eBook. The book is mostly on Amazon so physical books can be ordered for a book signing at an event. I have it priced just a bit above cost.
One key with editing is that you need someone willing to put in the time to do a careful job. In my experience, asking a co-worker tends not to work well. About the only time I get carefully edited (mostly copyediting) is when I'm actually working with an editor.
Rich is relative - I've written two books for a publisher and a co-authorship of a self-published book. In both cases, it was nowhere enough to recover the time spent on it. However, seeing the sales figures, seeing your work cited in publications and the reactions you get from people when they say "Oh you wrote that?" - that is another feeling by itself.
You have to remember that a lot of these books are written as groundwork for future plans - far easier to establish credentials saying "yeah I've written that book, you may have heard of it".
It also depends on what's your goal. In my case, I was just looking for a way to pay my bills and not having to go back to a job again. My monthly expenses where I live is around $150 and I'm earning about $200 per month as a self-published author. That's more than I need and I'm grateful for that.
- I published my first book after 8 months of intense writing and editing. I wrote everything from scratch. For the format, I first started with markdown and converted to pdf. I later realized the latex formula is not well supported so I rewrote in Latex. I outsourced other tasks on fiverr: proofreading & grammar checking and editing (about $180), latex touch-up and restructure about $100, book cover $100. I self-published on Amazon and got 60% royalty. It's too early to say but I wouldn't be surprised if I can get 12k for the first year. In terms of time spent, I would probably better off do something else for better ROI.
- I'm working on another book with Packt. They contacted me and the royalty is slightly higher (16%). They tend to push authors to meet deadlines etc. I still have no idea who the editors are and their background.
So if you're passionate about something and you already have followers, you'd be better off with self-publishing.
I greatly welcome your comment - thank you. I just wanted to point out that the proofreading might have come in a bit cheap: the following sentence on the Amazon summary has grammar errors (missing plurals).
> Data scientist, software engineer or data engineer who have a background in Machine Learning but never work on Machine Learning at scale will find this book helpful.
Interesting to see the process and royalty rates from a big (tech) publisher.
I wrote and self-published a book (really a web app structured like a book) on Ruby on Rails to Gumroad earlier this year and so far I have made ~$4,000 after Gumroad's small cut. Based on the royalty rates quoted here, I expect self-publishing is probably going to produce more raw dollars for most people, even those like me with no name recognition or built-in distribution channel for their book.
Of course, even with more money in self-publishing, the hourly rate I've made from writing is a small fraction of what I would have made filling that time with consulting work instead. I expect that in the long run, going with a traditional publisher would be more valuable just from the prestige attached to publishing a "real" book rather than going the self-publishing route. I've gotten a few consulting leads from my book, but a self-published book carries very little weight with most folks.
Author of "Hacking Healthcare" here. Myself and Fred were able to contribute to the discussions on our cover selection. I suspect some animals are probably much more in demand than others. Our editor was Andy Oram and he was outstanding. He is no longer with O'Reilly unfortunately.
Having worked with different publishers in different contexts, 90% of your experience will be as a result of your relationship with your editor.
I get asked sometimes by new authors, should I write for X or Y or Z. My answer is to always try and see who your editor will be and how well you think you can work with them.
I've written couple of books for a publisher (Practical Docker with Python - first & second editions - published by Apress) as well a co-authored a self published book - The CDK Book.
Apress has similar royalty tiers (10 to 20% of revenue, based on books sold) and I did get decent amount of copy edit support.
With the co-authorship, despite being four of us, some things slipped through the cracks. That said, with self publishing we were able to get it corrected and shipped - we maintain/ed bug reports, feedback in GitHub issues, and even the book build process is done using GitHub actions.
As I said elsewhere, the monetary aspect usually isn't the goal, you're far likely to spend more money in working on the book than recover it but it does give a nice boost if you wish to go the consulting route in terms of establishing authority/credibility.
I reviewed an oreilly book.
The process was well run, and most of my suggestions were acted upon by the editors/authors. Everyone was quite professional.
Compensation was de minmus, plus a copy of the book.
I strongly suspect that tech books do not sell as well as they did 25 years ago, and most tech books make little if any profit. There's an advantage to the publisher in having a wide variety of titles for a subscription service (I loved having a safari subscription), but, at least for me, the days of buying a giant and expensive tech book for just a few relevant chapters are over.
So the addressible market is smaller, and compensation is generally lower, though there are still indirect rewards
I strongly suspect that tech books do not sell as well as they did 25 years ago
I would think that goes without saying, tho I have no data.
though there are still indirect rewards
Yeah..."wrote an O'Reilly book on X" is a good way to move to the front of the interview line when X is a requirement. I've interviewed a couple of O'Reilly authors and currently work with one. Regardless of their other virtues and vices, they knew their topic.
I've had a fantastic time writing books (Python Workout and Pandas Workout) for Manning. They do great work, and are incredibly patient with me. They have helped to make my books far better than they otherwise would have been. I have no doubt that O'Reilly does a similarly good job. (I don't publish with O'Reilly, but I do teach for them online, and have been impressed.)
That said: Writing a book is a long slog. It's hard and frustrating. And at the end of the process, you have an amazing feeling and some money -- definitely less than you would have gotten if you had done consulting during that time, though.
I do corporate Python training. Do companies hire me because I've written books? Not really. But it definitely adds to my credibility.
At the end of the day, I write because I enjoy explaining and teaching, knowing I'm helping people all over the world, and seeing my name on a book cover. If you're doing it for the money, you're almost certainly going to be disappointed. Knowing that I'm working with great professionals who share my goals and want to create the best possible product adds to my motivation.
They aren't always great books. They're better now, but I remember over 25 years ago I bought "DNS & BIND" (Bat Book). I had to read the whole thing cover-to-cover three times before it began to make sense. It was so full of forward references that I just couldn't comprehend it, and I'm no stranger to technical content.
I used to love O'Reilly books, many years ago, and in my experience cookbooks and pocket reference guides were always a clear skip.
But nowadays I thought they've started distancing themselves from book publishing, considering that they don't have an obvious link to books on their homepage. They've been trying to position themselves as a learning platform for some time now.
I also (tried) to learn BIND via the bat book many years ago. I found it very hard to follow, and I was still early in my career so I just assumed that BIND was over my head and gave up.
To this day, DNS is a very aggravating topic for me. I wonder if much of that is due to the bad experience I had when I first tried to learn about it?
They are rarely good books these days. The fact the Python for Devops has 4.6 rating on Amazon baffles me. The book is unreadable, full of anecdotes that don't add up to anything. And as always, they gotta teach me python once again in the first few dozen pages.
I keep asking myself "what is so brilliant about this book that it deservers 4.6 stars?". Maybe the average developer can't tell bad content from decent.
I wrote a kids programming book for Bloomsbury and it was a nightmare. I wrote in org-mode, converted to docx and then sent off the copy. It was then commented on in Word, which I didn't own, so I had a weird libre-office back to org-mode workflow that was just painful. Then proofs were sent through as PDFs that inevitable messed up simple things like indentation, which I then had to fix, via comments on the PDFs.
> I wrote in org-mode, converted to docx and then sent off the copy
Did you expect them to redline it in something other than word if you sent a docx file?
All traditional (fiction/non-fiction) book publishing houses will use Word redlining. It is the defacto standard in the industry. Traditional publishing has 0 interest in switching to new software as far as I know.
Technical book publishers, and some university presses, will allow you a lot more leeway, but even some of those will ultimately redline on Word.
>The biggest issue I have in editing is that the official language of the book is American English, and I'm British, so I'm having to constantly correct my spellings. Changing 's' for 'z', removing silent 'u's, that sort of thing.
So no copy editing then? Thats the literal job of the copy editor, the author shouldnt be thinking about this at all. what are o'reilly doing to justify their 90% cut?
I've been approached by a few publishers but the math on revenues doesn't make any sense. If I could get myself to write a technical book I would 100% self publish.
Some of my peers accepted with the same publishers and now they can boast of being book authors for some online clout or some extra linkedin post.
It can be useful if you're building an online audience you can convert to money. I'd love to do that but I can't stomach the thought and I'm not willing to put effort into that.
In terms of CV it won't serve you to get a raise or get into places that pay more (for that, just leetcode). In this regards it is similar to doing talks and blogging.
It may be useful if you're applying not into software development but into a software evangelist role.
John Resig (inventor of Jquery) has written books across a variety of publishers. He has also written extensively about how he fared financially. These are worthy counter points to the discussions here about "only 10%."
> I have no idea, and neither does anyone else. The editing team have no say in which animal goes with which book. Aparently there's a team somewhere in O'Reilly whose job is to make the Animal selection. It'll be determined close to release date - some time next year.
15 years of wondering answered in a single blog post.
This is how it works currently due to the volume of manuscripts they are working on at any given time.
My editor told me that you used to be able to ask or try to make a case for a specific animal.
If you are writing a book about something that has an official animal or something, I am sure you can have some sway, but for the long tail of titles, you get one assigned randomly at some point.
If you are planning to publish a book, they are great to work with since they work with more engineer authors than other publishers may have experience with. A big part of it are the editors you work with there.
> If you are writing a book about something that has an official animal or something, I am sure you can have some sway, but for the long tail of titles, you get one assigned randomly at some point.
Wouldn't that potentially be a copyright violation? E.g. using elephant for PHP could be seen as trying to copy PHP's logo? [1] (Not a lawyer, this is just a question)
[1] This is likely a terrible example, as I don't think elephant is PHP's official logo, just an associated mascot like Java's Duke? Still... I don't know if this can be copyrighted.
And assuming PHP trademarked it's elephant, it has to be THE PHP elephant to be a violation, not an O'Reilly drawing elephant. Duke is a bad example, too because there's not real Duke to make a drawing of.
Incidentally, an elephant was used in their first big Hadoop book (and I would find it hard to believe that was coincidence): https://www.oreilly.com/library/view/hadoop-the-definitive/9.... There's a Hadoop-the-elephant logo that's trademarked by Apache, but elephant puns and even other drawings of yellow elephants abound in that community, and I'm not sure that's the kind of thing that trademark law could be enforced against even if the ASF wanted to.
My anecdotal experience: I didn't care for the first book, then my kid was doing a school report on a Hawaiian bird and asked if I can use that bird on a cover. I asked O'Reilly and they said yes, the animal wasn't taken. I guess chances are higher when the animal is more obscure.
That's the question for O'Reilly, but now how does Manning decide what goes on their covers (that's the only other publisher that has somewhat interesting ones... the Packt ones look more vanilla to me). Maybe a Manning author here would interject and give us some insight into that?
It's prestigious, and arguably the value of the prestige is worth more than the projected cash.
How much money is there in "Game Programming for Haskell", really? You release with O'Reilly and you get 10% or $1,000. You self-publish and you make more per sale, much more, but make much fewer sales - so, let's say $3,000.
But publishing with O'Reilly, you get accepted on the speaker's circuit and start getting asked to do contract work; publishing on your own, you don't (as much). It's arguably worth losing out on a one-time gain of $2,000 for that.
I haven't published with O'Reilly but I have published a couple editions of a book with Apress. Yeah, you don't write a technical book (or tech-adjacent) book for the money. But my observation is that a lot of people seem to be more impressed by a book for a known publisher than something you do through Kindle Direct Publishing.
I'm not sure that attitude is deserved. There are pros and cons with going through a publisher and am very unlikely to do it again. (Reputational benefits are pretty irrelevant to me at this point.) But no one should go in thinking it's about the money.)
Nobody is getting rich off writing a tech book. At some point a 10% or 50% royalty on a niche book for some random technology doesn’t really matter.
People author these books to increase the value of their personal “brand” or out of a passion for whatever technology they write about. If they are hoping to get rich or get a “livable” wage… they are going to be very disappointed.
* You won't make that much money directly, but I have probably made a good amount indirectly. The credibility from having published an O'Reilly book is not to be underestimated.
* I got the animal I wanted (honey badger) by asking for it. I did justify the ask by showing my editor the Bitcoin honeybadger memes.
* The book writing process was a grind. Because my book was so focused on programming exercises, I wrote my own suite of testing tools to make sure everything was consistent. It felt more like a software engineering project than a bunch of blog posts.
* I have since introduced more people to my intake editor and they've published their own O'Reilly books.
* I don't get much of a discount on my own book. I just buy them through Amazon.
* They give you a framed picture of your cover when your book publishes.