the ultimate insiderbook.sgit.ai v0.1.117 Star this site on GitHub

Research · leanpub · team/conductor/research/leanpub/04-what-the-best-leanpub-books-do.md

What the best Leanpub books do

Researcher: sub-agent for the conductor. All pages read on 30 October 2026 unless a row says otherwise.

How this was done, and its limits

1. Platform facts that shape every row

FactSource
Authors set a minimum and a suggested price; the reader picks a price between them (and may pay above the suggestion)https://leanpub.com/ansible-for-devops (page shows "Pick your price", Minimum and Suggested)
"All readers get free updates, regardless of when they bought the book or how much they paid"text on https://leanpub.com/datastyle and https://leanpub.com/os
A free minimum is allowed and common among the biggest books (see table)the book pages below
Authors can sell packages (book alone, book plus datasets and code, book plus lecture videos) and bundles of several bookshttps://leanpub.com/rprogramming ; https://leanpub.com/ansible-for-devops
Every page offers a free sample (PDF/EPUB), a table of contents, page and word counts, and a reader counthttps://leanpub.com/ansible-for-devops
A Reader Membership lets a member take a book for 1 "book credit" (a credit is worth $25) and about 2,000 books are free with membershiphttps://leanpub.com/leanpub_at_sixteen (2026)
Leanpub says it pays 80% royalties and keeps about $3 of a $20 salehttps://leanpub.com/the_most_painful_compliment (2026)
Leanpub's own account of its growth: an "organic flywheel" from readers to authors, now adding paid events, bootcamps and adshttps://leanpub.com/leanpub_at_sixteen ; https://leanpub.com/the_most_painful_compliment
Bestseller lists rank by revenue and copies in the past seven days, or by lifetime revenue and copieshttps://help.leanpub.com/en/articles/5395545-how-are-books-ranked-in-leanpub-s-bestseller-lists (search summary only)

2. The books (twelve, from the lifetime list)

Ordered by readers count shown, highest first; "n/s" is "not shown". Every row's source is the book's own page unless stated.

#Book, author, URLTopicReadersPrice modelState and last updateHow the page sellsAuthor's strategy outside Leanpub
1OpenIntro Statistics, David Diez, Mine Cetinkaya-Rundel, Christopher Barr and OpenIntro. https://leanpub.com/osCollege statistics textbook544,007Minimum free, suggested $15. Packages: the book, and the book plus tablet-friendly and past editions (1st to 4th), both free. Also 1 credit with membership100% complete; updated 2024-12-30; 422 pagesArgues against "expensive alternatives": free before the course starts, free forever including future editions, Amazon paperback at $25; claims about 20,000 students a year; says Leanpub revenue funds a US non-profit; sample download; page for teachersNon-profit with its own site (openintro.org) carrying free videos, labs, teacher registration and free desk copies; print on Amazon
2Introduction to Data Science, Rafael A. Irizarry. https://leanpub.com/datasciencebookR, data wrangling, machine learning195,514Minimum free, suggested $49.99; PDF; 1 translation (Spanish)n/sLong, concrete blurb of skills covered (R, dplyr, ggplot2, Git, knitr); full table of contents; author bio (Harvard professor)Academic with a university audience; page links a podcast episode (no other detail on the page)
3The Elements of Data Analytic Style, Jeff Leek. https://leanpub.com/datastyleData analysis practice136,067Minimum free, suggested $10; 31 translationsn/sShort; says it grew from the author's blog posts and tutorials; credits his co-direction of the Johns Hopkins Data Science Specialization, which "has enrolled more than 1.76 million people"Blog (Simply Statistics), a very large online course; the book is a companion to the course
4Ansible for DevOps, Jeff Geerling. https://leanpub.com/ansible-for-devopsServer automation115,385Minimum and suggested both $9.99 (no discount below, no ask above). Bundles: Vagrant Ansible $26.99, Ansible Bundle $29.97, Automated Servers $44.99. Also 1 credit100% complete; updated 2025-05-25; 470 pages, 91,580 wordsSubtitle "Server and configuration management for humans"; "tested with the latest version"; full table of contents and sample; podcast episodeHis own post of 19 Oct 2015 (https://www.jeffgeerling.com/blog/2015/finished-writing-my-first-book/) says almost two years of writing, "publishing-as-I-write", over 2,200 copies sold on Leanpub before the first edition, then Amazon and iTunes. Leanpub's help page cites him as a typical Lean Publishing success (search summary only)
5The Node Beginner Book, Manuel Kiessling. https://leanpub.com/nodebeginnerNode.js tutorial30,821$7.99 minimum and suggested; bundle "Javascript on Server Side" $11.99100% complete; updated 2018-11-13; 85 pages, 20,674 wordsAlmost no blurb ("No Description Available" on the list); sells on being short, cheap and a clear titlen/s on the page. A Leanpub testimonial attributed to the author says "over 28,000" sales after four years (search summary of https://leanpub.com/testimonials ; unchecked)
6Software Architecture for Developers, Simon Brown. https://leanpub.com/software-architecture-for-developersArchitecture27,883Minimum $20, suggested $25; 1 credit100% complete; updated 2022-05-28; 106 pagesSubtitle "Technical leadership and the balance with agility"; sample and table of contents; page count shownn/s on the page
7Introducing EventStorming, Alberto Brandolini. https://leanpub.com/introducing_eventstormingA workshop method12,327Minimum $9.99, suggested $19.9970% complete; updated 2021-08-26; 416 pages; author's note: "still in progress, after all these years""Straight from the inventor"; the author is the source of the methodn/s on the page
8The Hundred-Page Machine Learning Book, Andriy Burkov. https://leanpub.com/theMLbookMachine learningn/sMinimum $20, suggested $40100% complete; updated 2019-07-05Three named endorsements from well-known authors and professors above the blurb; the promise is in the title (100 pages); free sample; mentions a companion wikin/s on the page
9R Programming for Data Science, Roger D. Peng. https://leanpub.com/rprogrammingRn/sMinimum free, suggested $20. Three packages: book ($20); book plus datasets and R code ($25); book plus lecture videos, datasets and code (minimum $30, suggested $35)n/sBased on a university specialisation; print copies sold through LuluCourse (Johns Hopkins), print on Lulu, videos under a Creative Commons licence; bio lists ten other books
10Build APIs You Won't Hate, Phil Sturgeon. https://leanpub.com/build-apis-you-wont-hateAPI design5,236Minimum $26.99, suggested $29.99; a "cause gets" share; bundle "Vagrant API's" $38.99; 4 translations100% complete; updated 2016-03-30; 190 pagesBlurb speaks to the reader's pain ("Tasked with building an API for your company..."); samplen/s on the page
11Laravel: Code Bright, Dayle Rees. https://leanpub.com/codebrightLaravel PHP4,879Minimum $29.99, suggested $39.99; bundle "Code Evolution" $49.99; 7 translationsPublished while in progress (says so on the page); 449 pagesQuotes from the framework's own author and readers; the author explains why he raised the price from his earlier, cheaper bookn/s on the page
12The Art of Data Science, Roger D. Peng and Elizabeth Matsui. https://leanpub.com/artofdatascienceData analysis processn/sHeadline minimum $20, suggested $30; packages (book; book plus lecture videos). The package prices are laid out ambiguously on the page, so check before quotingn/sAimed at practitioners and managers; print on LuluAs above

Two outside case studies, both from the author's own words:

CaseWhat it saysSource and date
Nicholas Zakas, The Principles of Object-Oriented Programming in JavaScriptMinimum $14.99, suggested $19.99. Of 888 purchases, 243 paid the suggested price and 74 paid more; the average price was $16.74 (higher than the minimum); the highest was $78.62; one reader returned the book to buy it again for more. Total $14,866. Fewer copies than a publisher would sell, similar royalty income. The ebook was then picked up by a traditional publisher for a print edition. He wrote in Markdown, fixed typos at once from reader reports, and used his blog and Twitter for marketinghttps://humanwhocodes.com/blog/2014/03/18/leanpub-one-year-later (dated 18 March 2014 in its address)
Swizec Teller, an unfinished bookA draft of about 6,000 words with most chapters as outlines earned $959.88 in three days ($730.40 royalty); about 8,000 visitors came from Hacker News and the promo code was used 240 times; the full text was also on GitHubhttps://swizec.com/blog/i-made-dollar730-by-selling-an-unfinished-book-for-3-days/ (search summary only)
Jeff Geerling, free booksReported 328 paid sales a month before, 1,910 copies in two free months, then back to the old rate in May (search summary only; his post at the usual address now returns "not found", so check)https://www.jeffgeerling.com/blog/2015/finished-writing-my-first-book/ links to it
Ryan BiggCompares a traditional publisher keeping over two-thirds of revenue with Leanpub's cut; Leanpub author page says his books sold over 20,000 copies (search summary only; title of the 20,000 not confirmed)https://ryanbigg.com/2015/08/my-self-publishing-success-story

3. What the best ones have in common

  1. The biggest reader counts belong to books with a free minimum. The top four counts by a wide margin: 544,007 (free minimum), 195,514 (free), 136,067 (free), then 115,385 for a $9.99 book. Books with minimums of $20 or more show 5,000 to 28,000 readers. Caution: this compares readers, not money, and it includes books from large institutions. The sample is small and nothing here proves that free caused the reach.
  2. Most of the largest free books sit inside an audience engine. Leek and Peng have a university course with 1.76 million enrolled; Irizarry is a Harvard professor; OpenIntro is a non-profit with a free companion site for teachers. The book is the free front door to something larger.
  3. Finished, maintained books keep selling. Ansible was updated in May 2025 (ten years after its first edition) and has 115,385 readers; OpenIntro was updated in December 2024. Others were finished and left alone (Node Beginner, 2018) and still show tens of thousands.
  4. The pages are plain. Title, one-line subtitle, a short blurb, free sample, table of contents, page and word counts, author bio. Three pages use social proof (endorsements on the ML book; quotes on Code Bright; "straight from the inventor"; the student numbers on OpenIntro).
  5. Short and promise-in-the-title wins. 85 pages (Node Beginner), 100 pages (the ML book), 106 pages (Brown).
  6. Packages, bundles, translations and print are used by the larger books. Peng's three packages put videos and code on a paid ladder above a free book; Geerling has three bundles; Leek has 31 translations; Peng, Matsui and OpenIntro point to print copies (Lulu, Amazon). Zakas's ebook led to a print edition from a publisher.
  7. Pay what you want works at the top end too. In Zakas's data, the average was above the minimum, 74 of 888 paid more than the suggestion, and buyers who wanted to pay more did so.
  8. A launch needs a traffic source. Swizec's three days were Hacker News and Reddit; Zakas used his blog and Twitter. The platform did not bring them.
  9. Leanpub is moving towards memberships and events. Its own 2026 posts say the platform's growth was slow and organic, and that it is now building credits ($25 each), memberships and paid workshops. A free-minimum book has no price for a credit to discount, but being "free with membership" or listed on the store may still matter. This is unverified for our case.

4. Testing against the editor's pricing reasoning

The editor's reasoningWhat the evidence says
A $0 minimum, because an emailable, updatable reader is worth more than noneSupported by reach (top four reader counts) and by Leanpub's promise of free updates to everyone. Not yet shown: whether Leanpub lets an author email free readers. Each page has "Contact the author" but I found no page that says the author may email all readers. Check the author help pages.
Information should be shared; someone who cannot afford it should still have itMatches OpenIntro's stated mission and its "every student can access it, free forever" pitch.
Fewer paid sales, many more readers, a haloReach: yes, on the biggest books. Halo to other products: seen where there is a course, a companion site or a print edition (Leek, Peng, OpenIntro, Zakas's print edition). Not seen: any case of a book alone with no such engine. Geerling's reported free promotion faded after the free weeks (search summary only).
Readers come back to pay more or buy the next versionZakas shows buyers paying above the minimum and one buying again for more. Suggested prices work: Irizarry suggests $49.99 on a free minimum, Leek $10, Peng $20. How many pay is not shown on any page.

5. What this means for the book

Ranked, most to least certain:

  1. Keep the $0 minimum and set a suggested price (the pattern of the three largest books). Choose a suggested price that says the book has value, in the $10 to $20 range of the free-minimum books here, not the top of the range. Pay-what-you-want data (Zakas) says some readers pay above the minimum when they can.
  2. Tell readers in one sentence that all updates are free and arrive by email if Leanpub lets us email them; verify this first, because the editor's whole case rests on it. If it does not, the book's own site needs its own sign-up, and the page should link to it.
  3. Make the Leanpub page as plain as the best ones: a title with a one-line subtitle that carries the promise, a two-paragraph blurb, the free sample, the table of contents, the page and word counts, and a short bio. Put the strongest social proof (the talk, named people who will let us quote them) at the top, as the ML book does.
  4. Build or point to the audience engine. The big free books all sit on one. For us that is the talk, this repository and its site, and the interviews; the page should link to them, and they should link back. Do not expect Leanpub to supply readers.
  5. Publish in progress and update openly, and say so. Write "finished, with dated updates" or "in progress, updated on <date>" the way Brandolini and Rees do. Each update is a reason to email.
  6. Offer a paid ladder above the free book, only if there is something to put on it (Peng's datasets, code and videos; Geerling's bundles). For us the candidates are the transcripts, the record, and a printed edition; each needs the editor's decision.
  7. Plan for translations and print after the first release, as Leek (31 translations) and Zakas (print edition) did. Not before.
  8. Do not chase the weekly bestseller list. It ranks by revenue and copies sold in a week. A free-minimum book will not climb it, and the editor does not want that trade.
  9. Measure what the editor values: readers, emails collected, return visits, and how many pay above zero. Zakas published his numbers; ours should be published in the same plain way.

6. Open questions for a later pass

Where this came from

Leanpub pages read directly on 30 October 2026: https://leanpub.com/store/books?sort=bestsellers_lifetime ; the book pages named in the table (os, datasciencebook, datastyle, ansible-for-devops, nodebeginner, software-architecture-for-developers, introducing_eventstorming, theMLbook, rprogramming, build-apis-you-wont-hate, codebright, artofdatascience, ccie-spv4-comp-guide) ; https://leanpub.com/leanpub_at_sixteen ; https://leanpub.com/the_most_painful_compliment . Author posts read directly: https://www.jeffgeerling.com/blog/2015/finished-writing-my-first-book/ ; https://humanwhocodes.com/blog/2014/03/18/leanpub-one-year-later . Search summaries only (not opened): the Leanpub help articles on bestseller ranking, variable pricing and success stories; https://swizec.com/blog/i-made-dollar730-by-selling-an-unfinished-book-for-3-days/ ; https://ryanbigg.com/2015/08/my-self-publishing-success-story ; https://leanpub.com/testimonials . The editor's pricing reasoning is as given in the brief for this task. Every figure above should be copied into claims/ with its URL before the manuscript uses it.