Is my book ready to publish? A 10-point checklist for non-fiction

Ten questions test whether a non-fiction draft is ready, and three of them have a number to hit. The study drew them from the patterns in 2,529 editorial notes on 656 chapters. Run them in order: the first three test the book, the next four test each chapter's material, the last three test how chapters and the book end.

Write the book's controlling idea, model or contrast in one sentence, then check every chapter against it. All 14 books the study rated "very easy to follow" used one repeatable model or a concrete image per concept. A chapter that serves a second idea belongs in a second book or in the cut file.

Length tests the same thing from the outside. The median bestseller runs 91,574 words and the middle half runs 72,102 to 125,807. A quarter of the books are shorter and a quarter longer, so the band is a guide. Well below it, look for skipped steps; well above it, look for two books in one. Genre moves the band: see how long a non-fiction book should be.

Chapter one should open on a scene or story, and the final version should be at least a fifth shorter than your first draft of it. Twenty-three of 46 bestsellers open their first chapter on a scene or story, 11 on a bold claim and only one on a statistic.

First chapters are where good books over-reach. In the study, 72% of first chapters drew a note to tighten, against 46% of middle chapters. They also read no more easily than the rest of the book: the median difference was 0.1 Flesch points. The fix for an opening is cutting, not simplifying.

Every chapter should contain something you saw, did or got wrong yourself. Chapters rich in first-person experience drew "make it concrete" notes less than half as often as chapters without it: 17% against 38%. Your presence is what turns a principle into something a reader can picture.

The median book uses 11.5 first-person words per 1,000, and the middle half 6.2 to 20.0. Memoirs run near 45. If a chapter has none, add the moment you learned its point.

Search the draft for "over the next few years" and "eventually", and ask whether each one hides a scene. A turning point told in one sentence drew a "dramatize, don't summarize" note on 35% of chapters in the study, and on 53% of memoir chapters.

Scenes also help reading level, because they lean on plain, concrete words. Word length explains about 70% of the variation in Flesch scores across bestsellers, sentence length about 14%. Memoirs score 14 points above the median book on Flesch, mostly because of their shorter words. The bestseller band is grade 8 to 11; only 8 of 46 books read below grade 8. Check a passage with the readability checker.

Every "always", "never" and "studies show" needs a source and a stated limit. This is the most common note on great non-fiction: 53% of chapters drew a note to support a claim better, and 37% a note to qualify a sweeping one.

Evidence-led books drew support notes on 66% of chapters, against 36% for story-led ones. Survivorship bias was raised in 8 of 9 business books. More research does not cure it: chapters dense with research language drew more requests to qualify, not fewer. Size each claim to its source.

State the strongest objection to your argument in a form its holders would accept, then answer it. The study asked for this on 29% of chapters, as a note to engage objections or state the limits of the evidence.

A reader who disagrees with you is still reading. If the book never names their view, they stop trusting the rest. To find the objection, ask the person most likely to disagree for their best argument, in writing, and quote it back in their terms. One page that puts the other side well, and answers it, does more for your credibility than another stack of citations.

Replace a run of short examples with one case told in full, and include at least one that failed. "Use fewer, deeper cases" appeared on 24% of chapters overall and on 47% of business-book chapters, about twice the average.

A list of five companies proves less than one company followed from decision to result. A failure proves more than a success, because it shows the idea's limits. Keep the list to a sentence if you need it, and give the space to the case.

Each chapter should return, at its end, to the scene or question it opened with. "Pay off the opening" was the third most common note in the study, on 42% of chapters, and on 58% of self-help chapters. In 17 of 47 book reviews the verdict on structure was some version of "opens strongly, closes unevenly".

Length decides whether a reader gets to the payoff. A bestseller chapter is one sitting: the median book's chapters run 4,367 to 6,697 words in the middle half. Split any chapter that needs more than 40 minutes to read, about 9,500 words at the study's 238 words a minute. See how many words go in a chapter.

In a prescriptive book (self-help, business, faith practice), every chapter should end with the same named device that tells the reader what to do. Forty-one of 47 books in the study leave a clear next step, and 8 close every chapter with a named device.

Thinking, Fast and Slow ends chapters with "Speaking of…"; Mindset ends them with "Grow Your Mindset". Pick one form, name it, and use it in every chapter: three questions, one exercise, or a one-line rule to try this week. The six books with no next step are memoirs and histories, where that is a choice. If yours is a memoir, skip this check.

Your final chapter should resolve the story or question the book opened with, in a scene rather than a summary. Final chapters drew more "pay off the opening" notes than any other position: 46%, against 41% for first chapters and 38% for middle ones.

In 18 books, the note on the final chapter was about landing the ending. Read your first and last chapters back to back. If the last one only lists what the book said, it needs a scene that shows the question answered.

Yes, if you plan to sell the book. The checklist tells you whether your book is good enough to publish on paper: whether the ten features are there. An outside read tells you whether they work on a reader. It catches what you are too close to see: the step you skipped because you know it, the term you never defined, the chapter that loses a stranger.

The study shows the limit of numbers. Thinking, Fast and Slow and Sapiens both score 48 on Flesch, below the bestseller band, yet both were rated highly accessible. A formula cannot see a good explanation; a reader can.

You can skip a review for a family memoir that will not be sold, a collection of your blog posts, or a fifth book whose weak spots you already know. For anything else, try the stranger test first. Give a chapter to someone who has read none of the book, and watch where they hesitate or reread. Friends and family soften criticism, so choose a stranger to the material.

There are five ways to get a manuscript reviewed, from free to several thousand dollars. Most authors should start with the cheap, fast options and pay for a human edit only when they know what it is for. The full feedback ladder compares all five side by side.

• Self-edit with the checklist: free; seven passes, structure first. It is also how to edit a book for publishing before anyone else sees it.
• Beta readers: free through swaps; paid readers charge 0.5–0.8¢ a word at the EFA's 2026 rates, a median of $150 a project. A round takes weeks.
• The $19 Manuscript AI report: it reads the whole manuscript in one run and usually arrives in under 10 minutes, guaranteed within 2 hours. It answers eight editorial questions, three of which overlap this list (the main message, chapter openings and closings, the reader's next step). Each chapter gets a summary, recommendations, word count, reading time, Flesch score and grade, which you can set against the table above. See a sample report on Atomic Habits.
• An editorial assessment: an editor's letter on the whole book. Editors report a median of $600 per assessment (EFA, all lengths); for an 80,000-word book, per-word rates come to $1,200–2,000. Turnaround is set by each editor, so ask for it in the quote.
• A developmental edit: an editor works through structure and argument chapter by chapter. For an 80,000-word non-fiction book it costs $3,200–4,000 at the EFA's 2026 median rates, and $2,080–4,240 on Reedsy.

Use them in that order. Each step makes the next one cheaper, because the reader or editor spends their hours on what you could not fix yourself. Three mistakes waste any review: sending a first draft, choosing readers who only praise, and revising before you understand the notes.

The ten checks fit on one page. Print it, keep it next to the draft, and tick each one chapter by chapter.

1. One idea: the book's idea fits in one sentence, and every chapter serves it.
2. Opening: chapter one opens on a moment and is a fifth shorter than the first draft.
3. Presence: every chapter holds something you saw, did or got wrong.
4. Scenes: every "over the next few years" has been checked for a hidden scene.
5. Claims: every "always", "never" and "studies show" has a source and a limit.
6. Objections: the strongest objection is stated fairly and answered.
7. Cases: one developed case per chapter, including a failure, replaces the list.
8. Payoff: each chapter returns to its opening, and none needs more than 40 minutes.
9. Next step: in a prescriptive book, every chapter ends with the same named device.
10. Ending: the final chapter answers the opening question in a scene.

If you are submitting to a publisher, read what publishers look for in non-fiction manuscripts. For self-help, see how to improve a self-help book manuscript. To test reader reactions without waiting weeks, an AI beta reader is planned.