A book is about eighty thousand words. At four hundred words a session, two sessions a week, that's a hundred weeks.
Two years. Which means the project has to survive two Christmases, two summers, a house move or a job change or an illness, and roughly a hundred separate occasions on which you didn't feel like it.
Almost everything written about writing books addresses the sessions. The thing that determines whether it gets finished is the interruptions.
Why the daily-words advice misfires
Write every day is good advice for people whose job is writing. For someone doing it in evenings around other work, it fails in a specific way.
It sets a standard you'll break, and then breaking it becomes the event. Miss two days of an everyday commitment and you're not a person who missed two days — you're a person who broke the rule, and the rule was the thing holding it together.
It also can't accommodate the shape of an ordinary life. Some weeks have four available evenings and some have none, and a commitment that treats those identically will be wrong about half the time.
And it puts the emphasis on volume, when the thing that actually kills side-project books is loss of contact. A fortnight away from a manuscript doesn't cost you the words you'd have written. It costs you the state — where you were, what you'd decided, what the next scene or section was going to do — and rebuilding that is what makes the return feel enormous.
What survives two years
Four things, none of them about productivity.
A floor rather than a target. The commitment is the number you can hit on a bad week, not a good one. Two sessions, or even one, with everything above that optional. A floor doesn't produce failure, and failure is what ends these.
Sessions long enough to think. Ninety minutes is about right for drafting — long enough to get past the warm-up and hold a section in your head. Twenty minutes a day is excellent for keeping contact and poor at producing a book, and confusing the two is why people keep a daily habit for a month and have nothing.
A note at the end of every session, without exception. Where you stopped, what you were about to do. This is the single highest-value habit in a long project, because every gap you take is paid for out of it.
And stopping mid-thought. Finish a chapter and Thursday begins with deciding what the next chapter is. Stop halfway through a paragraph you know how to finish and Thursday begins with typing.
None of those four is about writing well, which is the point. Over a hundred weeks the quality of your prose is not the variable — you'll get better at that on the way. The variable is whether the project is still being touched in month eighteen, and everything above is aimed squarely at that.
The interruptions are the project
Two years guarantees them, so plan for them rather than around them.
The rule that matters: after two missed sessions, do a deliberately trivial one. Twenty minutes, two hundred words, or just reading the last page. Not to make progress — to stop the pattern converting into a decision, which is the mechanism that ends most of these.
For the longer breaks — a fortnight away, a hard month at work — park it explicitly, in writing, with a condition. Back to this after the move. A parked manuscript costs nothing to look at. An unparked one costs a small wince every time, and the wince accumulates into a reluctance to open the file at all.
And when you return, the first session is deliberately not writing. Read the last two thousand words, change nothing, add one sentence at the end. Twenty minutes, no judgement in it, and it converts the manuscript from a symbol back into a piece of work.
Where this is too comfortable
The honest objection, and it's a real one.
A floor of two sessions a week finishes a book in three years rather than two, and some books need finishing faster than that — a contract, a market, a subject that dates. If there's a real external deadline, this advice is too relaxed and you need a different arrangement, probably involving giving something else up.
There's also a version of the floor that becomes a hiding place. Two sessions a week, indefinitely, on a book that's been at forty thousand words for a year, is technically a live project and functionally a stalled one. The tell is the same as anywhere: whether the sessions move towards the difficult part or around it. A manuscript that's been revised from the start four times is not a manuscript being written slowly.
And some people genuinely do better with pressure than with floors. If you've finished things before under a deadline and never finished anything without one, believe your own evidence rather than this.
A hundred and four weeks
He starts in January. The plan is an hour a day, seven days.
It survives nineteen days. The first miss is a work trip, the second is being tired after the work trip, and by day twenty-five he hasn't written for a week and the daily commitment has become something he's failing at rather than doing.
The second attempt, in March, has a floor of two sessions and no daily target. Tuesday and Thursday, half seven to nine, plus whatever else happens.
Over the following two years the log has about two hundred and ten sessions. Roughly a third of the weeks have three or four; about fifteen weeks have none at all. There are two parked stretches — a five-week one for a house move and a three-week one when his father is ill — both with written conditions and notes about where he stopped.
The counter goes above fourteen days exactly twice, both times during the parked stretches, and both times the return costs one evening of reading rather than a fortnight of dread.
The book takes twenty-six months. The version with the daily target took nineteen days.
What the number is measuring
Not words. Days since you last touched it, which for a two-year project is the only measure that stays honest the whole way through.
Word count lies in both directions on a long book — it rises during a month of writing badly and falls during the best month you'll have, the one where you cut nine thousand words and fix the structure. The gap doesn't have an opinion about the quality of the work. It just tells you whether the thing is still alive.
Over a hundred weeks, that's the question that decides it.
Related: Ninety minutes is the unit · The two-day gap that decides everything · Why progress bars lie about knowledge work
Share: