Computing Applications Viewpoint

Do-It-Yourself Textbook Publishing

Comparing experiences publishing textbooks using traditional publishers and do-it-yourself methods.
  1. Introduction
  2. Challenges for Traditional Publishers
  3. Technical Challenges of DIY Publishing
  4. Economics of DIY Publishing
  5. Marketing of DIY Publishing
  6. Impact on Authors of DIY Publishing
  7. Conclusion
  8. References
  9. Authors
  10. Footnotes
  11. Figures
  12. Tables
Do-It-Yourself Textbook Publishing, illustration

We have just survived an adventure in do-it-yourself (DIY) publishing by finishing a software-engineering textbook. Our goal was to produce a high-quality, popular textbook that was inexpensive while still providing authors a decent royalty. We can tell you DIY publishing can lead to a highly rated, low-priced textbook. (See the figure comparing the ratings and prices of our DIY textbook with the three software engineering textbooks from traditional publishers published this decade).a Alas, as DIY marketing is still an open challenge, it is too early to know if DIY publishing can produce a popular textbook.b As one of us (Patterson) has coauthored five editions of two textbooks over the last 25 years with a traditional publisher,2,3 we can give potential authors a lay of the two lands.

Back to Top

Challenges for Traditional Publishers

The past 25 years have been difficult for publishers. The more efficient used-book market, the importing of low-cost books printed for overseas markets, and piracy have combined to significantly reduce sales of new books or new editions. Since most of the costs of book are the labor costs of development rather than the production costs of printing, the consequences have been to raise the prices of new books, to lower royalties to authors, and to force publishers to downsize. Obviously, higher prices make used books, imported foreign editions, and piracy even more attractive to less-scrupulous readers and resellers, creating a vicious circle. Less obviously, publishers have laid off copy editors, indexers, graphic artists, typesetters, and so forth, many of whom have become independent contractors that are then hired by traditional publishers on an as-needed basis. This independence makes them available to DIY publishers as well. Indeed, we invested about $10,000 to hire professionals to do all the work that we lacked the skills to do ourselves, such as cover design, graphic elements, and indexing.

Note that the outsourcing has also made traditional publishing more error prone and slower; authors have to be much more vigilant to prevent errors from being inserted during the distributed bookmaking process, and the time between when an author is done with the draft and the book appears has stretched to nine months!

Back to Top

Technical Challenges of DIY Publishing

We clearly want to be able to produce both an electronic book (ebook) and a print book. One of us (Fox) created an open source software pipelinec that can turn LaTeX into any ebook or print format. As long as authors are comfortable writing in LaTeX, the problem is now solved, although you must play with the text a bit to get figures and page breaks where you want them. While authors working with traditional publishers typically use simpler tools such as Microsoft Word, their text undergoes extensive human reprocessing, and it is distressingly easy for errors to creep in during the transcription process between file formats that many publishers now use. DIY authors must instead automate these tasks, and LaTeX is both up to the task and widely available to academic authors. There are many good choices for producing simple artwork; we used OmniGraffle for the figures we did ourselves.

While there are many options for ebooks, Amazon is the 800-pound gorilla of book selling. We tried several electronic portals with alpha and beta editions of our book,d but 99% of sales went through Amazon, so we decided to just go with the Kindle format for the first edition. Note that you do not need to buy a Kindle ereader to read the Kindle format; Amazon makes free versions of software readers that run on all laptops, tablets, and smartphones.

Back to Top

Economics of DIY Publishing

Author royalties from traditional publishers today are typically 10% to 15% of the net price of the book (what the book seller pays, not the list price). Indeed, the new ACM Books program offers authors a 10% royalty, which is the going rate today.

For DIY publishing, as long as you keep the price of the ebook below $10, Amazon offers authors 70% of the selling price of the book minus the downloading costs, which for our 5MB ebook was $0.75. For ebooks costing more than $10, the rate is 35%. Amazon is clearly encouraging prices for ebooks to be less than $10, as the royalty is lower for books between $10 and $20. We selected CreateSpace for print books, which is a Print-On-Demand (POD) service now owned by Amazon. We chose it over the longer-established Lulu because independent authors’ reviews of CreateSpace were generally more positive about customer service and turnaround time and because we assumed the connection to Amazon would result in quicker turnaround time until a finished draft was for sale on Amazon—it is 24 hours for CreateSpace. The royalty is a function of page count and list price. For our 500-page book printed in black-and-white with matte color covers, we get 35% of the difference between the book price and $6.75, which is the cost of the POD book. The good news is that Amazon lets us bundle the ebook and print book together at a single lower price in some countries in which they operate.

One benefit of DIY publishing is that we are able to keep the prices much lower than traditional textbooks and still get a decent royalty. We sell the ebook for $10 and the bundle of the print book and ebook for $40, making it a factor of 5 to 15 times cheaper than other software engineering books (see the figure).e To get about the same royalty per book under a traditional publishing model, we’d need to raise the price of the ebook to at least $40 and the price of the print book to at least $60, which presumably would still sell OK given the high prices of the traditionally published textbooks.

Some have argued for community-developed online textbooks. We think that it might work for advanced graduate textbooks, where the audience is more sophisticated, happy to read longer texts, and more forgiving.f We are skeptical it would work well for undergraduate textbooks, as it is important to have a clear, consistent perspective and vocabulary throughout the book. For undergraduates, what you omit is just as important as what you include: brevity is critical for today’s students, who have much less patience for reading than earlier generations, and it is difficult for many authors to be collectively brief.

One benefit of DIY publishing is that we are able to keep the prices much lower than traditional textbooks and still get a decent royalty.

Others have argued for free online textbooks.1 We think capitalism works; the royalty from a successful traditional textbook can be as much as half a professor’s salary. Making money gives authors a much stronger motivation to finish a book, keep the book up-to-date, and make constant improvements, which is critical in a fast moving field like ours. Indeed, we have updated the print book and ebook 12 times since January 2012. To put this into perspective, the experience of traditional publishers is that only 50% of authors who sign a contract ever complete a book, despite significant financial and legal incentives. Royalties also provide an income stream to employ others to help improve and complete the book. Our overall financial investment to produce and market the book was about $12,000, so doing it for free may not make sense.

Back to Top

Marketing of DIY Publishing

Traditional publishers maintain publicity mailing lists, set up booths at conferences and trade shows, and send salespeople to university campuses, all of which can usually be amortized across multiple books. We quickly learned the hard way that setting up tables at conferences does not amortize well if you have only one book to sell, as it is difficult to attract people. We also spent approximately $2,000 purchasing mailing lists, which once again has been outsourced and so now is available to DIY publishers. Fox had learned from his experience on the board of a community theater that a combination of email plus postcards works better than either one alone, so we had our designer create postcards to match the book’s look and feel and we did a combined postcard-plus-email campaign. Alas, this campaign had modest impact, so it is unlikely we will do it again.

In addition, as academics we were already being invited to give talks at other universities, which gave us opportunities to expose the book as well; we would usually travel with a few “comp” copies. Of course, “comp” copies for self-publishers means we pay for these out of pocket, but we gave away a great many, since our faculty colleagues are the decision makers and a physical copy of a book on your desk is more difficult to ignore.

Although we did not know it when we started the book (mid-2011), we were about to be offered the chance to adapt the first half of the campus course to a MOOC (Massive Open Online Course). MOOCs turned out to play a major role in the textbook’s development. We accelerated our writing in order to have an “alpha edition” consisting of half of the content, that is, the chapters that would be most useful to the MOOC students. Indeed, based on advice from colleagues who had offered MOOCs, we were already structuring the MOOC as short video segments interspersed with self-check questions and assignments; we decided to mirror that structure in the book, with each section in a chapter mapping to a topical segment in the MOOC. While the book was only recommended and not required for the MOOC, the MOOC was instrumental in increasing the book’s visibility. It also gave us class testing on steroids, as we got bug reports and comments from thousands of MOOC learners. Clearly, the MOOC helps with marketing, since faculty and practitioners enroll in MOOCs and they supply reviews on Amazon.

We have one cautionary tale about Amazon reviews. When we released the Alpha Edition, it was priced even lower than the current First Edition because more than one-third of the content had not yet been written (we wanted to release something in time for our MOOC, to get feedback from those students as well as our on-campus students). Despite repeated and prominent warnings in the book and in the Amazon book description about this fact, many readers gave us low reviews because content was missing. Based on reader feedback, we later decided to change the book’s title, which also required changing the ISBN number and opening a new Amazon product listing. This switch “broke the chain” of reviews of previous editions and wiped the slate clean. This turned out well, since the vastly improved Beta edition received 4.5 stars out of 5 from Amazon readers, a high review level that has continued into the First Edition, while our main established competitors have far lower Amazon reviews (see the figure). For better or worse, even people who purchase in brick-and-mortar stores rely on Amazon readers’ reviews, so it was good to start from a clean slate after extensive changes. So one lesson is to change the ISBN number when transitioning out of an Alpha Edition.

We quickly learned the hard way that setting up tables at conferences does not amortize well if you have only one book to sell.

We have learned two things through our “marketing” efforts. First, textbooks require domain-specific marketing: you have to identify and reach the very small set of readers who might be interested, so “mass” marketing techniques do not necessarily apply. This is relevant because many of the “marketing aids” provided by the Kindle author-facing portal are targeted at mass-market books and to authors who are likely to write more if successful, such as novelists. Second, in the past, publishers were the main source of information about new books, so your competition was similar titles from your publisher or other publishers; today, the competition has expanded to include the wealth of free information available online, an enormous used-book market, and so on, so the build-up may be slower. Indeed, the “Coca-Cola model” in which one brand dominates a field may not be an option for new arrivals.

Back to Top

Impact on Authors of DIY Publishing

As long as you do not mind writing in LaTeX, which admittedly is almost as much like programming as it is like writing, DIY publishing is wonderful for authors. The time between when we are done with a draft and people can buy the book is one day, not nine months. We took advantage of this flexibility to make a series of alpha and beta versions of our book for class testing.

Moreover, when we find errors, we can immediately update all ebooks already in the field and we can immediately correct all future ebooks and print books. POD also means there is no warehouse of books that must be depleted before we can bring out a new edition, as is the case for most traditional publishers. Flexibility in correcting errors and bringing out new editions is attractive for any fast moving field, but it is critical in a software-intensive textbook like ours, since new releases of software on which the book relies can be incompatible with the book. Indeed, the First Edition was published in March 2014, and a new release from Heroku in May 2014 already requires changes to the “book-ware” appendix. Such software velocity is inconsistent with a nine-month gap between what authors write and when it appears in print.

Back to Top


The resources are available today to produce a highly rated textbook, to do it more quickly than when working with traditional publishers, and to offer it at a price that is an order of magnitude lower. Given the marketing challenges, it is less obvious whether the book will be as popular as if we had gone with a traditional publisher, although a MOOC certainly helps. We may know in a few years if we are successful as DIY publishers; if the book never becomes popular, we may never know. But we are glad we went with DIY publishing, and we suspect others would be as well.

Back to Top

Back to Top

Back to Top

Back to Top


UF1 Figure. Ratings and prices on Amazon.com as of July 2014. The number of reviews are 2 (Braude and Bernstein), 38 (Fox and Patterson), 28 (Pressman), and 22 (Sommerville). Note the print+ebook bundle of Fox and Patterson costs the same as the print book only ($40), and that Braude and Bernstein have no ebook.

Back to Top


UT1 Table. Summary of self-publishing experience.

Back to top

    1. Arpaci-Dusseau, R. The case for free online books (FOBs): Experiences with Operating Systems: Three Easy Pieces; http://from-a-to-remzi.blogspot.com/2014/01/the-case-for-free-online-books-fobs.html.

    2. Hennessy, J. and Patterson, D. Computer Architecture, 5th Edition: A Quantitative Approach, 2012.

    3. Patterson, D. and Hennessy, J. Computer Organization and Design 5th Edition: The Hardware/Software Interface, 2014.

    a. In May 2014, our text Engineering Software as a Service: An Agile Approach Using Cloud Computing was rated 4.5 out of 5 stars on Amazon.com, while the most popular competing books are rated 2.0 stars (Software Engineering: A Practitioner's Approach and Software Engineering: Modern Approaches) to 2.9 stars (Software Engineering).

    b. There is a longer discussion of the DIY publishing in the online paper "Should You Self-Publish a Textbook? Should Anybody?"; http://www.saasbook.info/about.

    c. See http://bit.ly/1swgEsC. There is also a description of the pipeline in the online paper mentioned previously in footnote b.

    d. We tried Amazon Kindle, the Apple iBooks, and the Barnes and Noble Nook device, which uses Epub format. We wanted to try using Google Play to distribute Epub, which is watermarked PDF, but its baffling user interface thwarted us.

    e. In July 2014 Amazon sold the print version of Software Engineering: A Practitioner's Approach for $202 and the electronic version for $134, or a total of $336 for both. The total for Software Engineering is $260 or $120 for the ebook and $140 for the print book.

    f. For example, see Software Foundations by Pierce, B.C., Casinghino, C., Greenberg, M., Sjoberg, V., and Yorgey, B. (2010); http://www.cis.upenn.edu/~bcpierce/sf/.

Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More