How to Write a Book: A 10-Step Guide From First Idea to Finished Draft

TL;DR: Learning how to write a book is less about talent than about getting through ten predictable stages without quitting at any of them: test the idea, define the reader, choose a structure, outline enough to start, set a word-count plan, draft without editing, keep track of your facts, step away, revise in passes, and test the book on readers. Most books die between 30,000 and 60,000 words, when the manuscript becomes too big to hold in your head. That is exactly where StoryWith AI helps. It reads your whole manuscript, finds what is structurally wrong, ranks the fixes into one queue, and tests the result on readers, while never writing a sentence you will publish. Free plan with unlimited books, paid from $19/mo, ten languages.

---

Why most people who start a book never finish it

Almost everyone has a book in them. Very few people have a finished manuscript on their hard drive. The gap between the two is not a mystery, and it is not about whether you are "a real writer".

It comes down to three things.

The idea was never tested. Plenty of ideas that feel like novels are really short stories. Plenty of ideas that feel like business books are really one good article. You find out at chapter six, when there is nothing left to say.

The writer edits while drafting. Every session starts by rereading and polishing yesterday's pages. After three months, the first chapter is beautiful and there are only four chapters.

The book outgrows the writer's memory. Somewhere past 30,000 words, you can no longer see the whole thing at once. You forget what a character already knows, you repeat an argument, the middle starts to sag, and you cannot tell why. This is the point where most drafts quietly get abandoned.

The ten steps below are designed around those three failure points. Follow them in order the first time. Once you have finished one book, you will know which ones you can bend.

---

Step 1: Test the idea before you commit a year to it

Before you write a single chapter, find out whether your idea can carry a whole book. A novel usually runs 70,000 to 100,000 words. A nonfiction book typically runs 40,000 to 70,000. That is a lot of pages for one idea to fill.

Try writing your book in three sentences:

1. Who is this about, or who is it for?
2. What do they want, or what problem does the book solve?
3. What stands in the way, and what changes by the end?

If you can write those three sentences, you have a premise. If you cannot write the third one, you probably have a situation rather than a story, or a topic rather than an argument. That is fixable, but fix it now, not at chapter twelve.

A second test: list ten things that could happen in the book, or ten chapters it could contain. If you run dry at four, the idea needs more conflict, more complications or a bigger promise to the reader.

How StoryWith AI helps: you can start a project from one messy sentence. Capture the idea and the studio helps you shape it into a premise, a cast or a thesis, and a rough outline. It pressure-tests the idea by asking the awkward questions (what changes, what is at stake, what the reader is promised) without writing the book for you.

---

Step 2: Decide who the book is for

"Everyone" is not a reader. The single most useful decision you can make early is a clear picture of the one person you are writing for.

For fiction, that means a genre and a reader who loves it. A cosy mystery reader expects a puzzle, a community and no graphic violence. A thriller reader expects rising stakes and short chapters. A romance reader expects the relationship to be the main plot, with an emotionally satisfying ending. Breaking genre promises is allowed, but you should know which promises you are breaking.

For nonfiction, it means knowing exactly what your reader knows already, what they are struggling with, and what they will be able to do after reading. A book for first-time managers is a different book from one for experienced executives, even if the topic is the same.

Write a short description of your reader and pin it above your desk. Every time you are unsure whether a scene or section belongs, ask whether that person would miss it.

How StoryWith AI helps: fiction and nonfiction are separate paths from the start, because they are different jobs. A novel gets characters, scenes and a timeline. A memoir, guide or business book gets chapters, claims and the promise made to the reader. You never have to squeeze a self-help book into three acts.

---

Step 3: Choose a structure that fits your book

Structure is not a cage. It is the shape that tells the reader where they are and makes the ending feel earned. You do not need a complicated system, but you do need one.

Common fiction structures:

• Three acts. Setup, confrontation, resolution. The workhorse of most novels and films.
• The hero's journey. A character leaves the ordinary world, is tested, and returns changed. Popular in fantasy and adventure.
• A beat sheet. A list of specific turning points at rough percentages through the book, such as the inciting incident at around 10% and the midpoint twist at 50%.
• Genre-specific shapes. Romance has its own beats. Mysteries have a clue and suspect layout. Thrillers have ticking clocks.

Common nonfiction structures:

• Problem, solution, proof. Name the reader's problem, give the method, show it working.
• Step by step. Each chapter is one stage of a process, like this guide.
• Chronological. Common for memoir and history.
• Thematic. Each chapter explores one idea that builds toward a larger argument.

Pick one and move on. You can change it in revision. What you cannot do is revise a book that does not exist yet.

---

Step 4: Outline just enough to start

Writers argue endlessly about whether to plan everything (plotters) or discover the story as they go (pantsers). The honest answer is that most finished books come from somewhere in between.

A practical middle path is the one-line outline. Write one sentence for every chapter or scene you can currently see. Leave gaps where you do not know yet. For a novel, that might be thirty lines. For a nonfiction book, it might be twelve chapter titles with a sentence under each describing what the reader learns.

This gives you three things:

• A reason to sit down each day, because you know what the next scene is.
• A way to spot big structural problems early, such as a second half with no events in it.
• Permission to change your mind, because one line is cheap to delete.

If you are a natural discovery writer, outline only the next three scenes at a time. That is still enough to stop the blank page from winning.

---

Step 5: Set a word-count plan you can actually keep

This is where writing a book stops being a dream and becomes arithmetic. Do the maths once and you will never wonder whether you are "making progress" again.

Five hundred words is roughly a page and a half, and most people can write it in 30 to 45 minutes once they are warmed up. At that pace, a full novel draft takes about eight months. A shorter nonfiction book takes less.

Three rules make the plan survive real life:

1. Protect a time, not a mood. "Every weekday at 6:30 am" beats "whenever I feel inspired". Inspiration turns up more often when it knows where to find you.
2. Set a floor, not just a target. On a terrible day, 100 words still counts. Keeping the streak alive matters more than the number.
3. Track it. A simple spreadsheet or calendar with a mark for each session is enough. Watching the total rise is a surprisingly strong motivator.

---

Step 6: Write the first draft without editing it

The first draft has one job: to exist. It is allowed to be clumsy, repetitive and full of notes to yourself. Your job during drafting is to move forward, not to make anything good.

Some habits that keep a draft moving:

• Stop mid-sentence. End each session in the middle of a scene you know how to finish. Tomorrow you start by completing it, which is far easier than starting cold.
• Use placeholders. If you do not know a name, a date or a fact, type [CHECK] and keep going. Do the research later, in one batch.
• Leave notes, not fixes. When you realise chapter three needs changing, write a note at the top of your document. Do not go back and rewrite it.
• Skip the scene you are stuck on. Write the next one you can see. You can bridge the gap later, and often the solution becomes obvious once the surrounding scenes exist.

The single biggest mistake in learning how to write a book is rereading and polishing old chapters before the draft is done. It feels productive. It is actually the most common way books die.

---

Step 7: Keep track of your facts as the book grows

Every long book accumulates facts: names, ages, eye colours, places, dates, who knows what, and in nonfiction, every claim you have made and every source you promised to cite. At 10,000 words you can remember all of it. At 60,000 you cannot.

Traditionally, writers keep a story bible: a document or binder listing every character, place, rule and event, updated as they go. It works, but it is tedious to maintain, and most writers stop updating it around the point they need it most.

The problems it prevents are the ones readers notice instantly:

• A character refers to something they were never told.
• The timeline breaks, and a journey that took three days in chapter four takes one day in chapter nineteen.
• A nonfiction chapter relies on a concept you have not explained yet.
• Your argument in chapter eleven quietly contradicts chapter two.

How StoryWith AI helps: the story bible is built from what you actually wrote, not from a form you have to fill in, and it stays with the project between sessions, so you never have to re-explain your book. Its continuity checks read the whole manuscript and flag contradictions in facts, timelines and what each character knows. For nonfiction, fact-check and dependency scans flag claims that need a source, claims that contradict each other, and ideas that are used before they are introduced. Every finding links to the exact scene or section, so you can fix it in minutes.

---

Step 8: Finish, then step away

When you type the last line of the first draft, stop. Save a copy with the date in the file name, and do not open it for at least two weeks. Four is better.

This feels wasteful, but it is one of the most important steps in writing a book. While you are close to the draft, you read what you meant to write, not what is actually on the page. Distance lets you read it more like a stranger would.

Use the break to read books in your genre, fill in the research placeholders you left, or simply rest. You have just done something most people who start a book never do.

---

Step 9: Revise in passes, not all at once

Revision is where a draft becomes a book, and it is where many writers get overwhelmed. Reading a messy manuscript and trying to fix everything at once is exhausting and rarely works. The answer is to revise in passes, from the biggest problems to the smallest.

Pass 1: Structure. Does the book work as a whole? Is the beginning too slow? Does the middle sag? Does the ending pay off what the start promised? Are any scenes or chapters doing nothing? Fix these first, because there is no point polishing a chapter you are about to cut.

Pass 2: Characters or argument. In fiction, does each major character want something, change over the book, and sound like themselves? In nonfiction, does each chapter move the argument forward, or does it restate an earlier one? Is the evidence strong where it matters?

Pass 3: Scenes and sections. Does each scene have a purpose, a small change, and a reason to keep reading? Does each nonfiction section deliver what its heading promises?

Pass 4: Line editing. Only now do you work on sentences: rhythm, clarity, word choice, cutting repetition. This is the pass most people start with, and it is the one that matters least if the structure is broken.

Pass 5: Proofreading. Typos, punctuation and formatting, ideally done on a printed copy or a different screen, because your eye skips errors on a familiar layout.

The hard part is Pass 1 and Pass 2. They require seeing the whole book at once, which is exactly what the writer cannot do alone. A human developmental editor can, but a full developmental edit typically costs four figures and takes weeks.

How StoryWith AI helps: this is the stage the studio was built for. Import your draft from Word, PDF or plain text and it is analysed within minutes. For fiction, analysers map pacing and tension across the whole book and check scene purpose, character arcs, dialogue (does everyone sound the same?), style consistency and theme. For nonfiction, they check argument soundness, whether each chapter progresses or just restates an earlier one, and where the evidence thins out. You see the shape of your book as data instead of as a vague feeling that something is wrong.

Then comes the part that stops revision from being overwhelming. Every finding goes into one inbox, the Action Engine, which scores each problem on severity, confidence and effort. You open it and see your single best next move. Fix it, mark it done, and the queue re-ranks. A daunting rewrite becomes a list of evening-sized tasks, and the biggest problems always get solved first.

---

Step 10: Test the book on readers before you publish

You cannot read your own book as a stranger. You know what every sentence was supposed to mean, so you cannot tell where a reader gets confused, bored or lost. The only way to find out is to put the book in front of readers while you can still change it.

Traditionally, that means beta readers: people in your target audience who read the draft and tell you honestly what they felt. Good beta feedback answers questions like:

• Where did you stop wanting to read?
• Which character did you care about most, and least?
• Was anything confusing?
• In nonfiction: what did you learn, and where were you unconvinced?

The difficulty is getting enough readers, getting them to finish, and making sense of feedback that often contradicts itself.

How StoryWith AI helps: AI reader personas read your book as a defined kind of reader would, such as a genre fan, a sceptical expert or a complete beginner, and tell you where they would lose interest. Real beta readers get a private link and do not need to sign up, which is the difference between getting feedback and getting polite silence. The system then calibrates, comparing what the personas predicted with what real people said, so its read of your book gets more accurate over time. When you are ready to query agents or pitch, a readiness report and pitch kit help you present the book, and nonfiction authors can build a book proposal from the manuscript itself.

---

How StoryWith AI helps you write a book, at a glance

Most "AI for writers" tools do one thing: they generate text. That is the part of writing a book you least need help with, and the part that makes your book sound like everyone else's. StoryWith AI takes the opposite approach. It is a manuscript-aware studio: it reads your entire book and helps you make decisions about it, while the writing stays yours.

Why it deliberately does not write for you

StoryWith AI has no button that drafts chapter nineteen for you. That is not a missing feature. It is the main design decision behind the product, for three reasons.

Your voice is the product. Generated prose is the statistical average of everything already written in your genre. It is fluent and forgettable. Your voice is the only part of your book a reader cannot get anywhere else.

Your decisions are the book. Whether the sister forgives him, which chapter to cut, what your book is really about: those are the author's calls. Good software shows you the consequences of your choices. It should not make them.

Publishing platforms draw a line. Amazon KDP distinguishes AI-assisted work from AI-generated content, and agents, contests and retailers increasingly do the same. Using AI to analyse and improve your own writing is assisted work. Publishing prose a machine wrote has to be disclosed and is often treated differently.

StoryWith AI works alongside whatever you draft in, whether that is Word, Google Docs or Scrivener, in ten languages. There is a free plan with unlimited books and a monthly allowance of AI actions, and paid plans start at $19/mo.

---

A realistic timeline for your first book

If you write 500 words a day, five days a week, here is roughly what writing a book looks like:

• Weeks 1 to 2: Test the idea, define the reader, pick a structure and write a one-line outline.
• Weeks 3 to 34: Draft. About 2,500 words a week, 80,000 words in total.
• Weeks 35 to 38: Step away. Read, research and rest.
• Weeks 39 to 50: Revise in passes, starting with structure.
• Weeks 51 to 56: Beta readers, final revisions and proofreading.

That is just over a year for a full-length novel, and considerably less for a 50,000-word nonfiction book. Tools that show you what to fix first can cut the revision stage dramatically, because you stop spending weeks rereading the manuscript trying to work out what is wrong.

---

FAQ: how to write a book

How do I start writing a book with no experience? Start small and concrete. Write your idea in three sentences (who, what they want, what stands in the way), make a one-line outline of the scenes or chapters you can see, and commit to a small daily word count. Experience comes from finishing a draft, not from preparing to write one.

How long does it take to write a book? At 500 words a day, five days a week, an 80,000-word novel draft takes about eight months, and revision adds another three to six. Faster writers finish in months, and slower ones take a few years. Consistency matters far more than speed.

How many words is a book? Most adult novels run 70,000 to 100,000 words. Fantasy and science fiction often run longer, and middle-grade novels shorter. Nonfiction books typically run 40,000 to 70,000 words.

Should I outline my book or just write? Either works, but a light outline, one line per scene or chapter, helps most first-time writers finish. It gives every session a starting point without locking you into a plan you cannot change.

What is the hardest part of writing a book? For most writers it is the middle of the first draft and the first structural revision. Both require holding the whole book in your head, which gets harder as the manuscript grows. Whole-manuscript tools like StoryWith AI exist to take that burden off your memory.

Can AI write my book for me? Some tools will generate chapters, but the results tend to lose coherence over a full book, sound like the genre average, and must be disclosed as AI-generated on platforms such as Amazon KDP. Most authors get a better book by writing it themselves and using AI to analyse the manuscript, track continuity and prioritise revisions.

Do I need special software to write a book? No. You can draft a book in any word processor. Specialised tools earn their place later, when the manuscript is too big to manage by memory and you need to find structural problems, check continuity and decide what to fix first.

---

The short answer

How do you write a book? Test the idea, know your reader, pick a structure, outline just enough, do the word-count arithmetic, draft without editing, keep track of your facts, step away, revise from the biggest problems to the smallest, and test the book on real readers before you publish.

None of those steps requires genius. They require finishing each one before moving on, and having a way to see the whole book when it gets too big to see on your own. That is the job StoryWith AI was built to do.

Start free at storywith.ai. Unlimited books, import from Word or PDF, and your first full-manuscript analysis in minutes.