Fiction writing

Fiction writing splits into two activities that share a name and very little else. There is generating — producing new words that did not exist, which is the part people imagine and the part that feels like writing. And there is revising — changing what exists so that it does what you meant, which is where nearly all of the actual craft lives and which almost nobody pictures when they picture a writer.

The proportion between them surprises beginners. A finished piece of fiction is typically the product of far more revision time than drafting time, and the skills involved are close to opposite: drafting rewards momentum and suspended judgement, revising rewards patience and merciless reading. A great many people who describe themselves as blocked are trying to do both at once, which is a reliable way of doing neither.

The third thing, and the one that separates people who write from people who intend to write, is finishing. It is not a personality trait and it is not talent; it is a specific and learnable habit of getting to the end of a bad draft rather than restarting a good beginning. This page treats it as a skill, because it behaves like one.

What it costs to start

Nothing at allA notebook and a library card is a complete setup.

Where the craft lives

Revision, not draftingMost finished prose has been through several passes.

Typical novel length

70,000–120,000 wordsVaries substantially by genre, and shorter than most beginners assume.

The real bottleneck

Finishing thingsA learnable habit rather than a character trait.

A sequence that is much less romantic than the alternative and works far more often:

1. 1

   Step 1: Write something short first

   A complete story of two or three thousand words. It teaches beginning, middle and end in a fortnight rather than in two years, and finishing anything is the habit that everything else depends on.

   You will know it worked when: One finished thing, however bad.

2. 2

   Step 2: Finish the bad draft rather than perfecting the first chapter

   The single most common way a novel fails is a first chapter rewritten forty times. A complete bad draft can be fixed; a perfect opening attached to nothing cannot.

   You will know it worked when: An end, which is the only thing that can be revised.

3. 3

   Step 3: Put it away for a month

   Long enough to read it as a reader rather than as the person who wrote it. This is not a nicety — revision without distance mostly reproduces the same choices.

   You will know it worked when: The ability to see what is actually on the page.

4. 4

   Step 4: Revise in passes, one concern at a time

   Structure first, then scene-level work, then line editing, then proofreading. Fixing sentences in a chapter you will cut is the commonest waste of effort in the craft.

   You will know it worked when: A draft that improves rather than merely changes.

5. 5

   Step 5: Get it read by people who are not your friends

   A writing group, a workshop, or beta readers who read your genre. The useful feedback is where readers were bored or confused; the least useful is a list of what they would have written.

   You will know it worked when: Information about the reader’s experience.

6. 6

   Step 6: Start the next thing before you submit this one

   Both because the wait is long and because your second project is where the learning from the first one lands. Writers who submit and then wait tend to stop being writers for a year.

   You will know it worked when: A practice rather than a project.

On routine, the only defensible generalisation is that regular beats heroic. Whether that is five hundred words a day, an hour on Sunday or two thousand words in a burst once a fortnight matters much less than that it recurs, and prescriptions about writing every day come from people for whom that worked and are not a moral requirement. What is genuinely worth protecting is the ability to start quickly, and the most reliable trick for that — stopping mid-scene while you still know what comes next — costs nothing.

Rather than a method, here is the set of choices a piece of fiction actually requires, with the trade-offs stated. Most of them have no correct answer and all of them have consequences.

• Point of view — Who is telling this, and how much do they know? First person is intimate and constrained by what one character can see. Third limited is the workhorse of most contemporary fiction. Omniscient buys scope at the cost of closeness. The most common beginner error is not choosing badly but drifting between them within a scene, which reads as carelessness even to readers who cannot name what is wrong.
• Tense — Past is the default and is invisible to readers. Present is more immediate and more conspicuous, and a proportion of readers actively dislike it. Neither is better; the choice is about how much you want the telling noticed.
• Outlining or discovering — The genuine methodological fault line. Planning ahead trades spontaneity for structural soundness and a much lighter revision; discovering as you write trades a harder revision for material you would not have planned. Both produce good novels. Which one suits you is an empirical question about yourself, and the only way to answer it is to try both.
• Scene and summary — What gets dramatised in real time and what gets compressed into a sentence. Controlling this ratio is most of what people mean by pacing, and it is far more learnable than "pacing" makes it sound.
• How much to describe — Contemporary fiction generally describes less than fiction of a century ago, and genre conventions vary widely. The useful test is whether a detail is doing work — establishing character, creating unease, marking a change — rather than whether it is vivid.
• Dialogue — The part most beginners write best and most over-punctuate. Real speech transcribed is unreadable; fictional dialogue is a compressed impression of speech. Almost all attribution should be "said", which readers do not see, and adverbs attached to it are usually doing work the line itself should do.
• Structure — Something has to change, and the reader needs to feel the shape of the change. Beyond that, the available structures are numerous and the three-act model is a convention rather than a law — literary, episodic and non-linear forms all work and none of them is a failure to follow the rules.

Improvement in this craft is slower and less visible than in most, because the feedback loop is long and the standard keeps rising — you get better at noticing problems at roughly the rate you get better at avoiding them, so the work feels equally flawed for years. That is normal and it is the reason so many people conclude they are not improving when they are.

What actually moves the needle, in rough order of effect:

• Finishing more things — By a wide margin the biggest factor. Each completed piece teaches the whole arc; each abandoned one teaches only openings, which is why so many people have brilliant first chapters and no craft.
• Reading as a writer — Rereading something that worked and asking how. Where did the tension come from, what was withheld, how long was that scene actually. This is slow and it is the closest thing the craft has to an apprenticeship.
• Real critique — From readers who will tell you where they stopped caring. A group that only says what it liked is a pleasant social activity and does nothing for the work.
• Revising to a specific concern — A pass for structure, a pass for a single character’s consistency, a pass for over-explaining. General revision produces general improvement, which is to say very little.
• Reading your work aloud — The cheapest editing tool there is. Rhythm problems, repeated words and dialogue nobody would say are all audible and mostly invisible.
• Writing something outside your comfort — A different point of view, a different length, a genre you do not read. Uncomfortable and disproportionately instructive.

Writing is solitary and the culture around it is not. Local writing groups meet in libraries and pubs almost everywhere, and they are free or nearly so; workshops and courses run from a Saturday afternoon to a two-year degree; and there is an enormous online culture of critique communities, timed writing events and the annual November novel-drafting challenge, which has probably produced more finished first drafts than any other single institution in the hobby.

The one piece of etiquette worth learning early is how to receive critique, because it determines how much you get. Listen, do not defend the work in the room, and treat every reaction as accurate about the reader’s experience even where the suggested fix is wrong. A reader who was bored on page forty was genuinely bored on page forty; their theory about why may be useless and their observation is not.

Giving critique has its own convention: describe your experience rather than issue instructions, be specific about where rather than general about quality, and remember that "I would have written it differently" is not a fault in somebody else’s book. Groups that get this right are unreasonably valuable, and groups that get it wrong are worse than writing alone.

Fiction is not one thing with a length setting. The forms below differ in what they can do, how long the feedback loop is, and how quickly you find out whether you are any good — and the last of those matters more to a beginner than any of the craft differences.

• Flash fiction — Under about a thousand words. Finishes in an evening, which makes it the fastest way to practise endings — the part of storytelling almost nobody gets to rehearse. Compression is unforgiving and teaches more about sentences than any amount of drafting.
• Short stories — Roughly 1,500 to 7,500 words. The traditional apprenticeship form, with a live magazine market that accepts unsolicited work and returns a verdict in weeks rather than years. One complete arc, usually one central turn.
• Novelettes and novellas — From around 8,000 to 50,000 words. Genuinely awkward to place commercially and enjoying a revival in some genres. Long enough for a subplot, short enough to hold in your head at once, which many writers find is the ideal working length.
• The novel — 70,000 to 120,000 words for most adult fiction, varying substantially by genre. Everything about it is slower — the drafting, the revision, the submission — and the structural problems only become visible at a scale where fixing them is expensive.
• Serial fiction — Published in instalments as it is written, on a platform or by subscription. Enforces a deadline and a readership, at the cost of being unable to revise backwards. Closer to how a great deal of nineteenth-century fiction was actually produced.
• Linked collections — Short stories sharing a place, a cast or a preoccupation. Reads as a whole and is written in finishable pieces, which is a genuinely useful compromise for anybody who stalls in long projects.
• Interactive and branching fiction — Choice-based or parser-driven, with its own tools and its own annual competitions. Structure is the medium rather than a frame for it, and the craft is closer to game design than most fiction writers expect.

The practical advice that follows is the same one this page gives everywhere: the shorter forms are not a lesser version of the novel and they are not a warm-up. They complete, which means they teach the whole arc — beginning, middle and end — in a fortnight rather than in two years. A writer with ten finished short stories has practised finishing ten times; a writer two years into a first novel has practised it never.

Genre is a separate axis from length and is worth treating as one. It is a set of reader expectations and a place on a shelf rather than a quality judgement, the boundaries are drawn differently by different people, and plenty of the best fiction has been filed on both sides of them.

Most people who write fiction as a hobby never publish it, and that is a complete and legitimate version of the hobby rather than a failure to progress. For those who want readers, the routes are genuinely different from one another and choosing between them is a decision about what you want rather than a ladder.

• Short fiction to magazines — Submitting stories to literary or genre magazines. Free to enter, competitive, and the traditional way of building a track record. Rejection volume is high and impersonal, and this is where most writers learn not to take it personally.
• Competitions — Widely available and a reasonable way of getting a deadline and occasionally a readership. Entry fees are modest individually and mount up quickly, and the reputable ones are transparent about judging and prize money.
• Traditional publishing via an agent — For novels, the conventional route is to secure a literary agent, who then submits to publishers. Slow, competitive, and free — an agent is paid a percentage of what they earn you and never a fee in advance.
• Self-publishing — Entirely respectable and now a substantial part of the industry, particularly in some genres. You pay for what you use — editing, cover design, formatting — and you keep control and the rights. The costs are real and they are yours to choose.
• Small presses — Independent publishers, often specialising by genre or form, frequently open to direct submission without an agent. Advances are small or absent; the editorial relationship is often much closer.