Thoughts on writing and publishing Primer.

eeprimer (3)

I’m materially finished writing my 3rd book, The Engineering Executive’s Primer. There’s one last chapter to go through tech review, and a fine line editing pass, but the hard stuff is largely done. Of course, that’s an author’s perspective, there is other hard stuff still to be done by other folks in the process, particularly formatting and printing.

Each book is an education of sorts, and I decided to work with O’Reilly on this book to push myself on structure and consistency. I think both are significantly better in Primer than in my previous books, both of which I wrote, structured, and concept edited myself, although both benefited from others’ line editing. I’m confident it’s my best “book” yet, by which I specifically mean it’s the most bookish book I’ve written–the one that most closely approximates the Platonic form of a book. Each chapter is structured the same way. It deals with terms consistently. It doesn’t have chapters where I got annoyed and decided to leave them “as is” despite not wholly working.

This has been an important project to me, both because I believe it will do its small part in advancing the industry through supporting better engineering leadership, and also because it’s my first book since my stroke in early 2021. While objectively I’ve been wholly recovered for some time now, recovery from the unmeasurable is a subjective matter, and creating this book has been a meditation on the idea of “recovery.”

From here, I’ll ramble a bit on working with O’Reilly versus self-publishing, how I’m evaluating success for this book, marketing efforts, and what I’m thinking about working on next.

Working with O’Reilly

My experience working with O’Reilly, foremost with my editor, Virginia Wilson, was very positive. This certainly isn’t advice, as there’s too much context in each person’s individual situation, but at least you can get a sense of how I thought about it:

  1. In dogs, there’s this concept called “prey drive” which is when a dog’s predisposition to pursue a quarry to the extent that it ignores everything else. Sometimes that’s how I feel when I’m writing, but my target is finishing writing, and rather than running into a glass door, my risk is finishing something quite mediocre and deciding I’m done.

    Having an engaged concept editor who curbs my worst “it’s good enough as is” decisions was the most valuable part of working with O’Reilly. I also think it’s somewhat hard to self-edit on this, because it’s primarily an energy-level issue, but a recognition issue, e.g. I know the chapter isn’t particularly good, but at that moment I want to be done writing. There’s probably a structural solution to this, e.g. “never make a final decision on a chapter until I feel energized enough to write” but I don’t think I could write at my current pace, while simultaneously being challenged by my full-time work, if I didn’t push the boundaries at times.
  2. Working with O’Reilly really helped me in my progression as a book author versus a blog author. I’m continuing to work on improving my comfort with book-writing as distinct from blog-writing, and their feedback along the way has helped me considerably.

    Giving one concrete example, I often switch audiences across blog posts (e.g. one is written to an executive, another to a software engineer, a third to an engineering manager), which works in the context of a blog, but does not work in the context of a book
  3. For An Elegant Puzzle and Staff Engineer, I work with Nordlyset to handle publishing rights for non-English editions. This has been a vast improvement over managing the rights myself (which I did once for learning purposes), and I’ve been extremely happy working with Nordlyset. Working with O’Reilly gives me similar support for publishing rights within a single contract, which is quite nice. For me, this is neutral overall, and I will work with Nordlyset again if I publish future books with smaller publishers (or self-publish). However, I was only able to work with Nordlyset because I had already published a book with Stripe Press, if I’d only self-published I imagine I would still be handling these directly
  4. The timeline was very workable, and we came in under schedule. My guess is that I could have gotten the book out earlier if I’d self-published, but it would have been a worse book. I’m already comfortable writing to a schedule, so this was fine for me. I imagine it’s uncomfortable for folks who have not written this way before (e.g. first book and don’t have much experience writing with financial strings attached).

    At a few moments, I did miss the total timeline control that self-publishing provides. As mentioned earlier, I imagine I would have published a somewhat worse version of this book somewhat earlier if I’d self-published. Generally, I think that’s the wrong tradeoff to make for books on enduring topics (and in my opinion it doesn’t really make much sense to write books on non-enduring topics, when there are shorter formats to use instead), but I did miss the control. Particularly as someone who enjoys writing as a bit of “direct production” as distinct from the extremely team-heavy nature of my day-to-day leadership work. Maybe more importantly, I can imagine a counterfactual sequence of events where I simply couldn’t have published this book if I tried writing it during a busy period at work and didn’t have the option to drop the quality bar to release what was ready whenever my energy ran out on the project
  5. O’Reilly financial terms are meaningfully less generous than self-publishing. My guess is that I’ll make about one-half to one-third what I’d have made self-publishing. It’s true that there are fewer costs for me, but the upfront cost of self-publishing Staff Engineer was about $4,000 USD, which isn’t that much in the context of a book that you expect to sell a few 10,000 copies.

    For me, I have some indirect financial goals from writing, but not direct financial goals, so this wasn’t a sticking point. In general, I think optimizing for direct financial goals is the wrong choice for most first-time authors, and most repeat authors will have enough data to decide for themselves what the right tradeoff is for them 6. Probably the biggest open question for me is whether publishing with O’Reilly will help with distribution. My guess is that O’Reilly published Primer will sell less than self-published Staff Engineer, but largely due to different pools of potential readers. Lest this sound defeatist, I do think this book will be more impactful on the industry than my previous two books, which is my actual goal 7. I do think there’s some prestige value of publishing with a premier publisher like O’Reilly. Certainly it helps close the question of whether I self-published Staff Engineer (or any future books) voluntarily or if it was because they were simply unworthy of publishing, but that’s not a question I get much or really care about. I learned long ago that small groups on the internet will act as they please, and that it’s not really worth worrying about

Evaluating success

Some framing numbers as I think about success: My earlier books ended 2022 selling roughly 48,000 (Staff Engineer) and 60,000 (An Elegant Puzzle) copies. I’m guessing they’ll both reach about 70,000 copies by the end of 2023. That’s not a perfectly concrete number as it’s based on projecting a linear rate for Puzzle over the last 8 quarters and assuming that Staff Engineer’s sales on Amazon, ~16k copies since the end of 2022, hold somewhat consistent across audiobook and other channels. I’d be surprised if Primer sells quite as well as those books, which have broader audiences, but I didn’t expect either of those books to sell particularly well either, so we’ll all find out together.

Primer’s audience is anyone who aspires to be an engineering executive, works closely with an engineering executive, or is otherwise in an engineering leadership role. That is a sizable audience even if the narrowest interpretation (“only for engineering executives”) is a bit smaller.

Merging those ideas, numerically I’ll feel fine about Primer if it sells 10,000 copies in the first year (e.g. by the end of 2024). I’ll feel great about it if it sells 15,000 copies in the first year. While emotionally it’s hard not to get caught by sales numbers, intellectually I know that sales numbers are not a great way to evaluate books. Orienting around sales pulls you toward broad, generalized topics, which aren’t the topics I want to write about. What I’m most interested in is influencing executive leaders “doing the work” at real companies, and the meaningful work to be written there is in narrow topics.

Those narrow topics tend to be under-written about, often because publishers don’t think there’s an audience for the topic. This lack of competition allows you to write something niche, that also sells well, and is meaningfully influential on the industry. Staff Engineer is my personal example of this, which was at the time considered too narrow a topic to publish on before it proved out that market. Hopefully Primer does something similar.

In terms of evaluating whether Primer is doing the same, some rough thoughts:

  • An early sign that Staff Engineer was impacting industry was when it started getting referenced companies’ job descriptions. This was it directly impacting the vocabulary for Staff-plus roles
  • Another sign was when Staff Engineer began getting referenced by companies in their internal trainings, and in external conference talks
  • The influence became clear when folks in Staff-plus engineering roles started debating nuance and (dis)agreeing with some of the specific ideas. I increasingly believe that driving conversation is at least as interesting as driving consensus, and probably even moreso. Growth comes from encountering different perspectives and evolving our thinking through those encounters
  • Much like I’m easily distracted by sales numbers, I also love references in blog posts and on social media, but I think these are a bit second-order in terms of evaluating impact on the industry. I increasingly believe that only a small fraction of the folks I am trying to reach are active in those spaces, and meaningful signs need to exist beyond that ecosystem

None of these are particularly easy to instrument, and many are hard to detect unless someone drops me a note, but I think it’ll be clear in a year whether Primer is picking up that sort of engagement.

Marketing efforts

When working on my first book, Brianna Wolfson gave me some foundational advice on marketing books:

  1. Books are marketed over the first year they’re published, not just in the first week or month.
  2. Pre-existing distribution channels are much more reliable than anything you can create after publishing, including things like this blog’s mailing list, but in particular social media.
  3. Do three or four podcasts and talks.
  4. Don’t worry too much about doing more.

That advice still describes my approach, with a couple of additions of my own:

  1. Five, continuing to write is my best marketing strategy.
  2. Six, publishing stacks over time: each additional book you publish creates more visibility for all the previous books.

The most interesting advantages compound over time, and I’m pretty confident that’s the case for writing books, as long as you keep the quality bar consistently high and continue writing for generally the same audience.

What’s the next project?

I have two fairly concrete projects in mind to work on next. The first is something focused on engineering strategy, which is a topic I’ve just recently begun to feel I have a coherent argument to make around. The second is Infrastructure Engineering, which is a book I haven’t quite cracked how to write yet. I would be fairly surprised if my next writing project wasn’t one of these two, but it’ll be a bit. I’m writing one every two to three years at the current rate, so maybe 2026 or so!