The Economics of Writing a Technical Book

This content has been archived. It may no longer be accurate or relevant.

From Medium:

I am not an expert. I have co-authored a single book in 2017 called Cloud Native Infrastructure for O’Reilly Media. Many people have asked me what it was like so I will attempt to explain the process, time investment, and financial incentive here.

. . . .

The process was about what I expected. I was introduced to Brian, our first of three editors, from someone I knew who was already writing their third book. They thought I might be a good fit for what they were looking for so they made the introduction.

I thought about it for a couple weeks and then submitted a formal book proposal which entailed filling out a Word document template and emailing it to the editor. I didn’t hear back for about 3 weeks and then, after a follow-up email, heard the proposal was approved. After a kick-off call it was suggested that I find a co-author to help write the book. I had a week to find one and then needed to sign a contract with O’Reilly for dates and deliverables. I interviewed a few people and Kris Nova and I complemented each others skills perfectly for the content we wanted to cover in the book. She agreed it sounded like a good topic and she was excited to take on the challenge.

The contract seemed fairly standard and focused around content ownership and royalties split. The default split between authors is 50/50 which we stuck with. The contract stipulated that Kris and I own the copyright for the content, but O’Reilly has exclusive rights to use the content any way they see fit throughout the world now and in the future for the duration of the copyright.

Once the contract was signed there was a steady pace of work as we both figured out how to lay out content and what we should write about. O’Reilly provides a platform called Atlas for writing which is quite good. You write in plain text AsciiDoc and then O’Reilly’s Atlas platform can generate a PDF, or other formats, via the web interface or API. We both used atlas-cli to generate PDFs as we wrote. Generating the PDFs was a good feedback loop on the content. It helped make sure formatting was right and also allowed us to take a step back to read what we wrote.

. . . .

On March 1st we were assigned our cover animal which Kris and I named Andy O’Connor the Andean Condor. We were pretty excited to see the cover for the first time even if the subtitle went through multiple revisions. We didn’t get to pick the animal or the picture. We were told up front we wouldn’t get to pick the animal so we knew what to expect. We were also told that Tyrannosaurus Rex and unicorns are not allowed.

We kept writing until the 1/2 draft was due in early June. We turned it in and got less feedback than we expected, but it was still good to have a fresh set of eyes looking at it. We didn’t like what we had created. We had written almost 6 chapters and threw away 3 of them. The first two were heavily edited and the remaining chapter was trimmed down significantly and turned into an appendix.

We had some more planning meetings and came up with a revised outline that we submitted to our editor for review. By this time we were on our 2nd editor who wasn’t very familiar with the project so we got very little feedback and went with what we had.

. . . .

The first Tuesday of September the full draft was due and then went into a review process. There were technical reviewers we were able to suggest but mostly O’Reilly pulled from a pool of their trusted reviewers. We got minimal feedback from most of them (a survey form) and one returned notes on the PDF. We had a week to make edits. During this time the draft was made available as a preview on Safari books. In retrospect I wish we had posted preview chapters sooner which was something our first editor suggested, but we were both too embarrassed to follow through.

It wasn’t enough feedback for me so I reached out to more people and sent them chapters looking for someone to tell me it sucks and why. Luckily, I found someone who would give me the harsh feedback I wanted and I had about 3 days to incorporate their changes into the book before it went off to post production.

The last push was very difficult and stressful. There were a lot of big changes on the last weekend which was a risk, but I think in the end made the book better. The final weekend we moved some chapters around and wrote a chapter from scratch for content we felt was missing.

. . . .

I believe the first PDF came back with more than 1300 edits. Overall there were more than 2000 changes made during post. I later found out this amount of edits is fairly standard for our book length. We had about 3 weeks of emailing large, heavily notated PDFs back and forth which was no fun compared to the plain text git workflow of writing.

. . . .

All in all I worked from Feb — Oct for roughly 5 nights a week at 2–3 hours per night. I also worked about 3 weekends non-stop when a draft or final edits were due. Roughly I’d say I worked about 500 hours total. That was only my time and doesn’t include Kris’. I was lucky to have a co-author to share the load.

. . . .

At the end of final edits I was done (contractually and mentally). I had read through the entire book at least three times and much of the content was starting to lose meaning. After sending the final edited PDF I wanted to stress about missing an edit before going to bed, but I was too tired to care.

. . . .

O’Reilly provides an affiliate program which was terrible to set up and in the end hardly worth the time. You get a cut from all sales that go through your link but I have never received any money from affiliate book sales. The only money I got was when someone used my link and then bought a ticket to an O’Reilly conference.

. . . .

I attempted to set up an affiliate program for Amazon but my application was denied. Amazon offers an author central site to create a 1998 inspired author profile page and an out of date book sale statistics and rankings. I’m really not sure the point of creating the Amazon author information outside of claiming the book(s) you author and confirming that you have a terrible book rank.

. . . .

I would suggest anyone writing a book spend a night to register a domain and set one up. I launched it on August 31, 2017 and it has over 4,600 visits which is terrible by most website standards but good as a place to funnel users for info.

. . . .

From December through March the book has sold 1337 copies. I have no idea how well other books in this category sell. This total also includes 2 book signings at conferences that were sponsored by the CNCF (Thank you!) which was roughly 150 physical books total. On average, the book has sold 222 copies per month which is greatly skewed by the first month which had 930 sales. The last month (March) had 34 physical book sales. I suspect that number will go down even more over the next few months.

Sponsorships was an unexpected source of income. We have been lucky enough to have 3 sponsors so far. The sponsor pays O’Reilly for exclusive rights to provide a PDF and optional print version of the book. The company gets to put a forward in the book that Kris, me, and an O’Reilly editor approve. Once the sponsor completes their contract with O’Reilly they can do whatever they want with the books. Usually, the PDF gets put behind a web form so you fill out your email address and the company uses it for marketing services and getting customer leads. Physical books are usually given away at conferences where they can scan badges.

. . . .

Each full book sponsorship for one month nets me $3,705 and partial sponsorships give an amount based on percentage of the book sponsored (e.g. 5 chapters in a 10 chapter book is 50% sponsored). That’s much better than I expected because a one month full sponsorship is more than all other sales combined.

. . . .

There are also some other sponsorships that I think count as ebook sales but I never got a clear answer how royalties work for those. Book licensing incurs a small payment but I’m unclear how that is used. From my statements, three people have licensed the book or excerpts from it which has netted me $2.37.

. . . .

My April 2018 statement (sales from December — March) says I’ve made $11,554.15 which roughly breaks down to $23 per hour for the estimated 500 hours of work. Without the three sponsorships that would have been $5.29 per hour.

. . . .

The book has provided a few other opportunities that I probably wouldn’t have had. So far I’ve done a couple podcast interviews, spoken at a few events, did one webinar, and have had a few opportunities for more writing projects with O’Reilly (some of which I’ve taken).

Would I write another book? Not for the foreseeable future. I would like to update Cloud Native Infrastructure to keep it fresh with current industry trends, but another book from scratch is not a year long project I’d be looking forward to at this time.

Link to the rest at Medium