Outlining a nonfiction book so it holds together
Nonfiction rarely fails sentence by sentence. It fails structurally: chapters that repeat each other, a chapter four that assumes something first explained in chapter nine, a book that turns out to be seven articles about roughly the same subject stapled together.
Readers describe that as a pile of blog posts. They are describing an outline problem, and an outline problem is the cheapest kind to fix, because you can fix it before the words exist.
What follows is the nonfiction version of that work. Fiction has the same problem and different tools — Turning an idea into a book you can write is the companion to this one.
Start with the promise
Write one sentence naming what the reader can do or understand when they finish that they could not before.
Not a topic. A book about productivity is a topic and commits you to nothing. After this book you can plan a week so the important work happens before the urgent work is a promise: it names a reader, a change, and a scope. It also tells you what the book is not about, which is the useful half.
Everything in the book either serves that promise or comes out. That sounds obvious until you apply it to the chapter you were most looking forward to writing.
Keep the sentence where you can see it. Tightened, it is also most of your blurb and your description field — see How to write a blurb that sells the book — and it is the first thing a proposal has to state, which is why The nonfiction proposal, and why self-publishers should write one is worth writing even when nobody is going to read it.
Then pick a shape
Most nonfiction uses one of five, and choosing deliberately is what separates a book from a collection.
Problem, then solution. Establish that something is broken and costly, then spend the rest of the book fixing it. Strong pull. The trap is spending half the book on the problem; the reader agreed with you in chapter one.
Chronological. Time carries the reader — the obvious choice for history, memoir and anything where cause follows cause. The trap is that time is not always the interesting order: a year where nothing happened still gets a chapter.
A sequence of steps. Do this, then this, then this. The right shape for a how-to, and the easiest to test: try to do the thing using only your chapters, in order, and see where you get stuck. The trap is steps that are not really sequential, dressed up as though they are.
A set of principles. Independent ideas supporting one argument. Flexible, and the most dangerous, because nothing forces an order and nothing stops you adding a twelfth principle that is really the fourth one again. These need the chapter test below more than any other kind.
A journey or a case-study spine. One story, or several, carrying the argument. Readable, memorable, and the shape most likely to leave a reader unable to say what they should do on Monday. Pair it with something that states the point plainly.
Pick one and let it govern. A book that is chronological for six chapters and then becomes a set of principles has changed the rules on a reader who had learned them.
The chapter test
This is the tool that does the most work, and it takes an afternoon.
Give every chapter one sentence saying what it does for the reader. Not what it is about — what it does. "Shows why time-blocking fails for people with interrupt-driven jobs" is a job. "Interruptions" is a label on a folder.
Two things fall out immediately.
If two chapters have the same sentence, they are one chapter. This is the single most common structural fault in nonfiction, and it is invisible from inside the draft, because the two chapters contain different examples and feel different to write. The reader does not experience examples; they experience being told the same thing twice.
A chapter you cannot write a sentence for does not exist yet. It is a title, a folder of research, or an intention. That is fine at this stage — but mark it, because it will otherwise be the chapter you cannot finish in month four, and you will blame the writing. Keeping that folder in a form you can still use a year later is its own skill: Keeping track of your research.
The sentence is a working note, not the heading a reader sees. Turning it into a title that says what the chapter does without sounding like a slide is a later job — Chapter titles and headings in nonfiction.
In the Interior Formatter. Your outline becomes real structure: chapters detected from the manuscript even in a file with no styles, and the contents generated from those real chapter starts rather than typed by hand — so a contents page that reads badly is telling you something about the outline, not about the typesetting. Start your book → · about the Interior Formatter
Order by dependency, not by interest
For each chapter, ask one question: what must the reader already know for this to make sense?
Order the answers. That is the spine, and it is usually not the order you find most interesting. The temptation is to front-load your best material, which leaves a reader either lost or quietly re-reading chapter two to work out what a term meant.
Dependency also catches the forward reference — the paragraph that says as we will see in chapter nine. One or two are fine. A book full of them is a book whose chapters are in the wrong order and knows it.
Where the dependencies genuinely fork, you have a book with parts: a foundation everyone reads, then sections that can be taken in any order. Say so in the introduction. See Parts, sections and the pages between chapters.
The proportion check
Beside each chapter, write the length you expect in words. Then look at the spread.
A chapter that will run three times longer than its neighbors is nearly always two chapters that have not been separated yet. One that comes out at a quarter of the others is usually a section of the chapter before it.
This is not a rule about even lengths — some subjects are bigger. It is that a large gap is a reliable signal of a structure you have not finished thinking about. Splitting a chapter in an outline takes five minutes. Splitting it in a finished draft means rewriting both halves, re-cutting the transitions either side, and re-checking every cross-reference. For how the totals add up, see How long should a book be?.
Outlines change, and that is what they are for
Yours will. You will write chapter six and discover it belongs at three, or that two principles are one, or that the promise was slightly wrong.
An outline that survives the draft unchanged usually means you were not paying attention. Its job is not to prevent change — it is to make change visible, so moving a chapter is a decision you take once, deliberately, instead of a drift you notice only when a reader tells you the book felt like a pile of blog posts.
Make your book
The tools are free to use: name your book, bring your manuscript or your artwork, and work as long as you like. You only pay when you want the files without the preview watermark.
Next: Does your book need a table of contents? · Also: Does your nonfiction book need an index? · The editing passes that have to happen before typesetting