A project measured in months fails differently from an article. Nothing dramatic goes wrong. You lose track of what is finished, the schedule slips quietly, and one week without progress becomes six. This guide covers the machinery that prevents that: chapter states, a visible board, an external reader, and a plan for the middle.
- Keep one page of chapter lines and edit it as the shape moves.
- Allow four states only: planned, drafted, revised, done.
- Give every chapter a next action small enough to start today.
- Set a minimum you can hold in your worst week, not your best.
- Plan for the middle and for assembly, where long projects fail.
Structure at chapter level first
Before anything is drafted, write one line per chapter saying what it does for the reader and what has changed by the end of it. Ten or twelve lines on a single page. This is an argument about sequence, not a summary of content, and it is cheap to rearrange while it is still lines.
Keep that page open for the whole project and edit it whenever the shape moves. On a long project the chapter list is the artefact you will reread hundreds of times. When it stops matching the manuscript, you have lost the map at exactly the point you most need one.
Name the states, and allow only those
Four states cover a manuscript: planned, drafted, revised, done. Planned means the chapter has a line and some material. Drafted means prose exists end to end, however bad. Revised means it has been rebuilt once with structure in view. Done means it needs nothing but proofreading.
The discipline lies in refusing anything between them. Half-drafted is not a state; it is a chapter that is planned. Almost done is not a state either, and it is the one that hides months of work. Vague labels are how a project appears eighty per cent complete for a year.
Books are not abandoned in a moment of despair; they are abandoned by a fortnight nobody noticed.
One board for the whole manuscript
Put every chapter on a single page with its state, its current length, and the one next action it needs. That page is the dashboard, and updating it takes a minute at the end of a session. Nothing else should be required to know where the project stands.
The next-action column matters more than the numbers. Read a source, rewrite the opening, decide whether section three survives: something you could begin in ten minutes. A chapter with no next action recorded is a chapter you will avoid, because deciding is harder than writing.
Working to somebody else's schedule
Supervisors and editors are a resource with a queue. They read slowly, in batches, and rarely at the moment you would like. Plan around the delay. Agree what you will send, in what shape, and by when, and always have other work to move to while a chapter sits with them.
Send work at the stage where the response is worth having. A supervisor reading a rough draft comments on prose you were going to cut anyway; the same person reading a chapter plan can save you a month. Say plainly what kind of response you need this time.
A cadence you can hold in a bad month
Set a minimum you can hit in your worst week rather than your best one. A modest daily target that survives illness, teaching, and family will beat an ambitious one abandoned in February. Progress on a long project is mostly a function of weeks not missed.
Protect the same hours rather than the same word count, and log what happened each day in one line. Across months that log becomes evidence, which is useful when the project feels stalled and you genuinely cannot remember whether it actually is.
The middle, where books die
The middle is where the original idea has stopped being exciting and the ending is too far off to pull you. Nothing is wrong, which is the problem. There is no crisis to respond to, only a long stretch of ordinary work with no visible edge to it.
Three things help. Make progress visible, since a board showing four chapters drafted argues with the feeling that nothing is happening. Work on a chapter you want to write instead of the next one in order. Shorten the horizon to the current chapter, because a whole book is not a usable unit of attention.
Expect one structural revision around here, and treat it as an ordinary event rather than a verdict. Most long projects discover their real shape halfway through. Rewriting the chapter list at that point is progress, even though it moves no word count at all.
Assembly and the last stretch
A manuscript written chapter by chapter is not yet a book. Read it straight through in as few sittings as you can manage, marking repetition, contradictions, and terms defined twice. These faults are invisible while you work chapter by chapter and glaring in one pass.
Then finish the unglamorous list: references checked, figures numbered, cross-references corrected, formatting to whatever the institution or publisher demands. Schedule real time for it. Long projects overrun at the end because this work appears in no plan that treats writing as the whole job.
A chapter list before and after it became a status board
Ch 1 intro — mostly written? Ch 2 background — done I think, still needs the new references. Ch 3 methods — done. Ch 4 fieldwork — big mess, half of it is still notes, need to talk to J about scope. Ch 5 — this is the one I keep putting off. Ch 6 case studies — two written, third one maybe not needed? Ch 7 conclusion — nothing yet, obviously. Overall probably seventy per cent there. Need to speed up after Easter or this is going to slide again.
1 Introduction — planned, 900 words — next: rewrite once chapter 5 exists. 2 Background — drafted, 6,400 — next: add the four new references, then mark revised. 3 Methods — done, 5,100 — next: nothing. 4 Fieldwork — planned, 2,300 words of notes — next: ask J to confirm scope by 14 March. 5 Interpretation — planned, nothing written — next: one page of chapter lines, Tuesday morning. 6 Case studies — drafted, 7,800 — next: decide whether the third case survives. 7 Conclusion — planned, nothing written — next: leave alone until 5 is drafted. Two chapters drafted, one done, four planned. Nothing here is at seventy per cent.
The first list feels honest and cannot be acted on. The second says the same thing in states and next actions, and it reveals that the project is earlier than it felt.
Common mistakes
| The mistake | Why it costs you, and what to do instead |
|---|---|
| Measuring the book by word count alone | Words counted include material you will cut and exclude the thinking that makes a chapter work. Track states and next actions beside the totals, or a good month of deletion reads as going backwards. |
| Stopping while feedback is outstanding | A chapter sent to a supervisor or editor can sit for weeks. Queue other work before you send it, so the reply arrives as an improvement to make rather than as permission to start again. |
| Drafting chapters strictly in order | Order is a reader's requirement, not a writer's. Writing whichever chapter you can see clearly keeps momentum, and an introduction is almost always better written last, once the argument has stopped moving. |
| Reading the mid-project rewrite as failure | Discovering the real structure halfway through is what long projects do. Quitting there loses everything; rewriting the chapter list loses a week. Expect one such revision and it stops feeling terminal. |
Your checklist
- Write one line per chapter saying what it does for the reader.
- Mark every chapter planned, drafted, revised, or done.
- Give each chapter a next action you could start today.
- Agree with your supervisor or editor what goes out and when.
- Set a weekly minimum you can hit in a bad week.
- Book time for the read-through and the final mechanical checks.
Key terms
- Chapter state
- One of a small fixed set of labels describing how far a chapter has come, chosen so no chapter can sit in a vague condition such as almost finished.
- Next action
- The smallest task that would move a chapter forward, written down in advance so returning to the work requires no fresh decision.
- Read-through
- A pass in which the whole manuscript is read in sequence, in as few sittings as possible, to catch repetition and contradiction invisible chapter by chapter.
Setting a project up properly?
Frequently asked questions
How long should a chapter take?
+
Long enough that you stop measuring it that way. Chapters differ enormously: one summarises settled material, another carries the argument the whole project rests on. Plan by state and next action rather than by an average that describes none of them accurately.
What if my supervisor changes direction?
+
Get the change in writing, in your own words, and send it back for confirmation. Then update the chapter list before you touch any prose. Changes of direction are survivable; ambiguity about whether one actually happened is what costs months.
Should I revise as I go or draft everything first?
+
Draft a chapter, then move on, and resist the urge to rebuild it straight away. Structural revision is far cheaper once the neighbouring chapters exist, because most structural faults concern what a chapter assumes of the reader and what it repeats from somewhere else in the manuscript.
How do I return after months away?
+
Read the chapter list and the last few entries in your daily log before opening any prose. Then do one small next action instead of trying to reload the entire project. Re-entry fails when the first task waiting for you is large.
Is a thesis workflow different from a book?
+
The machinery is the same: chapter states, a board, a cadence, an external reader. The constraints differ. A thesis has a fixed format, an examiner, and a committee; a book has a market, an editor, and a catalogue slot that will not move.
What do I do about the chapter I dread?
+
Name the dread precisely. It is usually one unresolved decision rather than the whole chapter. Write that decision down as a question, answer it in a page of notes, and the chapter generally stops being the thing you are avoiding.