# The Leanpub years — the control group

**Source** the founder, brief 44, 28 August 2026.
**Status** SOURCE MATERIAL. This is evidence the book quotes, not prose the book uses. The
founder's words are quoted; the framing between them is the agent's and is marked as such.
**Why it exists** brief 42 asks part one to compare this workflow against the ordinary one.
This is the ordinary one, run by the same person, on the same kind of material, before the
machinery existed. It is the closest thing this book has to a control group.

---

## What was already good

The comparison is not "a good workflow against a bad one". It is against a workflow the
founder himself describes as *"a really powerful workflow"*:

> creating basically books from markdown that were integrated via GitHub through Dropbox,
> which actually allow me to reuse a lot of the blogs that I have written… some of the
> early versions of the source documents that you now have.

Markdown as the source. Version control. Reuse of existing writing. **Some of the same
source documents this estate carries today.** Anyone building the comparison should note
that the earlier setup already had three of the ideas this book argues for.

## What it achieved

> It was a success from the point of view that I was able to publish a lot of great ideas.
> I helped a lot of people, and it showed me that publishing books isn't that hard.

- Published, and read.
- Free with an optional payment, *"and some did, which is quite cool"*.
- A couple taken to print-ready PDF, onto KDP and Amazon — *"lessons that we're now
  using"*, and they are: the print PDF this estate builds descends from that experience.

## Where it stopped, and it was not the writing

This is the finding, and it is specific:

> What I couldn't do is get them to the actual level of quality required to have a full
> finished book, and to have a book that can reach a wider audience.

Writing the content worked. Writing it took real time and produced real volume. **The phase
after writing is where it failed:**

> Once I finished that, it's the next phase that becomes quite problematic. And it's the
> phase that we're doing right now, which is this ability to refactor, to change, to think
> about things, to take a step back, to challenge, to really question what we're doing.

## The sentence

> **You almost start to be locked by the first version of the content, because making
> changes becomes quite painful.**

*(Agent's note: this is the book's own argument stated as its opposite. Everything the
machinery does — the projection chain, the gates, the byte-identical rebuild, per-book
versioning — exists so that the first version does not lock anything.)*

## What made it possible to want that

> I don't have an emotional attachment to a lot of the stuff that gets created. I care
> about the final product. I care about the experience. I care about the message… that
> there's a conversion between what I'm trying to convey and how it comes out.

*(Agent's note: this belongs beside the caveat about programming experience. Refactoring a
book requires being willing to lose prose you wrote. That is a disposition, not a tool, and
a reader who lacks it will not get the benefit however good the machinery is.)*

## How they ended

> Those books ultimately would never finish because they ran out of time. And I ran out of
> a team almost to do it, and they never reached good enough quality that somebody wanted
> to invest in it — an editor, a publisher.

Two scarcities, and the book should be careful to distinguish them: **time**, and **a
team**. This estate's answer to the second is agents, and the making-of book is the argument
that they are a real answer. That claim should be made in the reader's hearing of this
story, not before it.

## How part one should use this

*(Agent's recommendation, for the editor and the writer.)*

1. **Do not open with it.** The reader has not yet been shown what is possible, and a
   failure story lands as an excuse before it lands as evidence.
2. **Use it as the hinge**, after the feature tour, to answer *why does any of this matter*.
3. **Keep the successes.** A control group that only failed is not credible, and this one
   did not only fail.
4. **Do not claim the machinery would have saved those books.** It is untested against
   them. The honest claim is narrower: the thing that stopped them is the thing this
   machinery is built for.